Search

In the Big Data era, context-aware mobile recommender systems are crucial in assisting citizens and tourists in making informed decisions, providing a suitable way for users to find the relevant data. These systems should be proactive, able to detect the ideal time and location to provide recommendations for a specific item or activity. To accomplish this, push-based recommender systems can be employed, utilizing context rules to determine when a recommendation should be initiated. However, there is very limited reported experience in defining and implementing such systems and a complete generic solution that adapts flexibly to the preferences of users and protects their privacy is still missing. In this paper, we present a novel approach where appropriate types of recommendations are provided automatically, without the need for user input. Our proposal allows users to easily activate, deactivate, customize, and create rules for improved personalization. Additionally, the module that, based on the context, decides the types of recommendations required is executed on the user's mobile device, reducing wireless communication and safeguarding the user's privacy, as context data are evaluated locally. To illustrate the approach, we have developed R-Rules, a prototype for Android devices focused on the triggering of recommendation rules, which provides a friendly user interface that facilitates user personalization. We have evaluated various technological options and demonstrated the feasibility, performance, and scalability of the proposal, as well as its suitability to users' needs. • Generic architecture for mobile push-based recommendations. • Decision module executing on the user device: privacy and wireless savings. • Flexible user-defined rules to decide when to fire a specific type of recommendation. • Android prototype and extensive experimental evaluation. [ABSTRACT FROM AUTHOR]

Highlights • This paper suggests two methods based on AI tools for medical image segmentation. • The first method uses CBR to guide a region-growing process for segmentation. • The second method based on Deep Learning performs segmentation with a small data set. • Both methods were tested on CT-scans images from children with tumoral kidneys. • They provided accurate enough results to consider an automatic. segmentation Abstract Most often, image segmentation is not fully automated and a user is required to lead the process in order to obtain correct results. In a medical context, segmentation can furnish much information to surgeons, but this task is rarely executed. Artificial Intelligence (AI) is a powerful approach for devising a viable solution to fully automated treatment. In this paper, we have focused on kidneys deformed by nephroblastoma. However, a frequent medical constraint is encountered which is a lack of sufficient data with which to train our system. In function of this constraint, two AI approaches were used to segment these structures. First, a Case Based Reasoning (CBR) approach was defined which can enhance the growth of regions for segmentation of deformed kidneys using an adaptation phase to modify coordinates of recovered seeds. This CBR approach was confronted with manual region growing and a Convolutional Neural Network (CNN). The CBR system succeeded in performing the best segmentation for the kidney with a mean Dice of 0.83. Deep Learning was then examined as a possible solution, using the latest performing networks for image segmentation. However, for relevant efficiency, this method requires a large data set. An option would be to manually segment only certain representative slices from a patient and then use them to train a Convolutional Neural Network (CNN) how to segment. In this article the authors propose an evaluation of a CNN for medical image segmentation following different training sets with a variable number of manual segmentations. To choose slices to train the CNN, an Overlearning Vector for Valid Sparse SegmentatIONs (OV 2 ASSION) was used, with the notion of gap between two slices from the training set. This protocol made it possible to obtain reliable segmentations of per patient with a small data set and to determine that only 26% of initial segmented slices are required to obtain a complete segmentation of a patient with a mean Dice of 0.897. [ABSTRACT FROM AUTHOR]

Highlights • The paper provides multi-view clustering algorithms for relational data. • The algorithms are based on Particle Swarm Optimization. • They are able to select the relevant views for the clustering task. • The algorithms consider eleven different fitness functions. • Experiments with real multi-view data sets shows their usefulness. Abstract Clustering of multi-view data has received increasing attention since it explores multiple views of data sets aiming at improving clustering accuracy. Particle Swarm Optimization (PSO) is a well-known population-based meta-heuristic successfully used in cluster analysis. This paper introduces two hybrid clustering methods for multi-view relational data. These hybrid methods combine PSO and hard clustering algorithms based on multiple dissimilarity matrices. These methods take advantage of the global convergence ability of PSO and the local exploitation of hard clustering algorithms in the position update step, aiming to improve the balance between exploitation and exploration processes. Moreover, the paper provides adapted versions of 11 fitness functions suitable for vector data aiming at dealing with multi-view relational data. Two performance criteria were used to evaluate the clustering quality using the two proposed methods over eleven real-world data sets including image and document data sets. Among new findings, it was observed that the top three fitness functions are Silhouette index, Xu index and Intra-cluster homogeneity. The performance of the proposed algorithms was compared with previous single and multi-view relational clustering algorithms. The results show that the proposed methods significantly outperformed the other algorithms in the majority of cases. The results reinforce the importance of the application of techniques such as PSO-based clustering algorithms in the field of expert systems and machine learning. Such application enhances classification accuracy and cluster compactness. Besides, the proposed algorithms can be useful tools in content-based image retrieval systems, providing good categorizations and automatically learning relevance weights for each cluster of images and sets of views. [ABSTRACT FROM AUTHOR]

Highlights • AMR Graphs are based upon PropBanks which limits them. • Deep Learning Models capture both the syntactic and semantic structure. • Requirement of large data set limits the use of Deep Learning Models. • Need of generalized framework for abstractive summaries is the need of time. Abstract Summarization, is to reduce the size of the document while preserving the meaning, is one of the most researched areas among the Natural Language Processing (NLP) community. Summarization techniques, on the basis of whether the exact sentences are considered as they appear in the original text or new sentences are generated using natural language processing techniques, are categorized into extractive and abstractive techniques. Extractive summarization has been a very extensively researched topic and has reached to its maturity stage. Now the research has shifted towards the abstractive summarization. The complexities underlying with the natural language text makes abstractive summarization a difficult and a challenging task. This paper presents a comprehensive review of the various works performed in abstractive summarization field. For this purpose, we have selected the recent papers on this topic from Elsevier, ACM, IEEE, Springer, ACL Anthology, Cornell University Library and Google Scholar. The papers are categorized according to the type of abstractive technique used. The paper lists down the various challenges and discusses the future direction for research in this field. Along with these, we have identified the advantages and disadvantages of various methods used for abstractive summarization. We have also listed down the various tools which have been used or developed by researchers for abstractive summarization. The paper also discusses the evaluation techniques being used for assessing the abstractive summaries. [ABSTRACT FROM AUTHOR]

Stochastic Gradient Descent (SGD) is one of the most popular first-order methods to solve optimization problems in large-scale, which has also been widely studied in intelligence system, deep neural networks, or machine learning. In this paper we set out to revisit SGD with practical concerns in mind and hope to provide intuition on how SGD should be done in the right way in applications for expert and intelligence systems. In literature, when implementing SGD, regularization is often added to the loss function, to avoid ill-conditioning or overfitting, or to obtain solutions with desirable sparse structure. Intuitively, regularized loss function deviates from the true loss and in this case SGD may lead to a suboptimal solution. In this paper we revisit the matter of l 2 regularization for SGD, to investigate whether or when an explicit regularizer is necessary to obtain the desirable performance. We further introduce a simple stochastic algorithm (ASG) using the accumulated stochastic gradient of the un-regularized loss. Then experiments are carried out on benchmark data sets to validate the theoretical analysis. The findings show that for l 2 regularization, (1) SGD without explicit regularizing (SGDE) actually possesses an implicit regularizer, and in the sense of upper bound of the convergence rate, it outperforms SGD regularized explicitly (SGDER) with a constant advantage; (2) ASG without explicit regularizing outperforms both SGDE and SGDER, especially in the case where the number of iteration T cannot be pre-specified; (3) for SGD algorithms, the schemes that places flexible weight on the output of the latest iteration can give a better trade-off, compared with the scheme using the output of the last iteration or taking the average of the output of each iteration, and this suggests that in application a tunable averaging scheme is preferable. This study provides insights on using SGD algorithms for big data applications, e.g., to accelerate SVMs or regularized regression in large-scale, or to improve the performance of online learning or real time forecasting (control). In particular, when T is pre-specified, SGDE (without an explicit regularizer) can give us a well enough perforce with simplicity and understandability; when T cannot be pre-specified, ASG can improve the performance of standard SGD algorithms. [ABSTRACT FROM AUTHOR]

In recent years, the growing availability of huge amounts of information, generated in every sector at high speed and in a wide variety of forms and formats, is unprecedented. The ability to harness big data is an opportunity to obtain more accurate analyses and to improve decision-making in industry, government and many other organizations. However, handling big data may be challenging and proper data integration is a key dimension in achieving high information quality. In this paper, we propose a novel approach to data integration that calibrates online generated big data with interview based customer survey data. A common issue of customer surveys is that responses are often overly positive, making it difficult to identify areas of weaknesses in organizations. On the other hand, online reviews are often overly negative, hampering an accurate evaluation of areas of excellence. The proposed methodology calibrates the levels of unbalanced responses in different data sources via resampling and performs data integration using Bayesian Networks to propagate the new re-balanced information. In this paper we show, with a case study example, how the novel data integration approach allows businesses and organizations to get a bias corrected appraisal of the level of satisfaction of their customers. The application is based on the integration of online data of review blogs and customer satisfaction surveys from the San Francisco airport. We illustrate how this integration enhances the information quality of the data analytic work in four of InfoQ dimensions, namely, Data Structure, Data Integration, Temporal Relevance and Chronology of Data and Goal. [ABSTRACT FROM AUTHOR]

In the era of remote sensing (RS) big data, in order to alleviate the time cost of acquiring RS images, recommending RS images that meet users' individual needs continues to be an urgent technology. However, the technology accomplished to date has two main problems: (1) they rely on the users' queries and thus lack initiative and cannot tap the users' potential interests and (2) they restrict the users' preferences to temporal and/or spatial information while ignoring other attributes and are not compatible with visual information. In an effort to fully explore the features of RS images and thereby achieve accurate active recommendations, in this paper we propose a new Multi-modal Knowledge graph-aware Deep Graph Attention Network (MMKDGAT) which we built upon graph convolutional networks. Specifically, we first constructed a multi-modal knowledge graph (MMKG) for RS images to integrate their various attributes as well as visual information, and then we conduct deep relational attention-based information aggregation to enrich the node representations with multi-modal information and higher-order collaborative signals. Our extensive experiments on two simulated RS image recommendation datasets demonstrated that our MMKDGAT achieved noticeable improvement over several state-of-the-art methods in so far as active recommendation accuracy and cold-start recommendation. • Fuse multi-modal information to enrich representations. • Use an attention mechanism to estimate the neighbor's importance. • Aggregate higher-order information in a light and effective way. • Construct two datasets and make them publicly available. • Significant improvement in accuracy over existing baselines. [ABSTRACT FROM AUTHOR]

In today's age of big data and information, large-scale group decision-making has become an essential aspect of modern economy, science, and technology. This paper proposes a large-scale group decision-making method that leverages group-oriented rough dominance relation to identify the worst group when addressing complex issues that involve a large number of decision-makers. The proposed method entails building a set-valued ordered information system that utilizes clustering learning to reduce data dimensions and reduce the dimensions of decision space, thereby improving the efficiency of the decision-making process. Additionally, it proposes a novel group-oriented rough dominance relation based on dominance-based rough set theory. By clarifying this advantage relationship, more targeted focus is placed on the group that needs improvement, thereby improving decision-making effectiveness. The proposed method calculates the advantage degree of alternative group plans to select the worst group. The main purpose is to compare the advantages and disadvantages of different groups by defining a metric that quantifies the rough dominance relation between groups, thereby improving the objectivity and repeatability of the decision-making process. Finally, the study applies the proposed method to a case study aimed at improving various types of services in European scenic spots, and the benefits of the method are discussed. Experimental analysis shows that the method in this study can screen the groups that should be improved most in this case, showing the benefits and applicability of the proposed method, and providing valuable insights for complex decision-making problems involving multiple decision-makers. • The clustering learning is used for dimensionality reduction. • Using the group-oriented rough dominance relation for decision-making aggregation. • A new method called LSGDM is designed. [ABSTRACT FROM AUTHOR]

Accurate and fast short-term load forecasting is crucial in efficiently managing energy production and distribution. As such, many different algorithms have been proposed to address this topic, including hybrid models that combine clustering with other forecasting techniques. One of these algorithms is bigPSF, an algorithm that combines K-means clustering and a similarity search optimized for its use in distributed environments. The work presented in this paper aims to improve the time required to execute the algorithm with two main contributions. First, some of the issues of the original proposal that limited the number of cores simultaneously used are studied and highlighted. Second, a version of the algorithm optimized for Graphics Processing Unit (GPU) is proposed, solving the previously mentioned issues while taking into account the GPU architecture and memory structure. Experimentation was done with seven years of real-world electric demand data from Uruguay. Results show that the proposed algorithm executed consistently faster than the original version, achieving speedups up to 500 times faster during the training phase. [ABSTRACT FROM AUTHOR]

In this paper, we focus on finding clusters in partially categorized data sets. We propose a semi-supervised version of Gaussian mixture model, called C3L , which retrieves natural subgroups of given categories. In contrast to other semi-supervised models, C3L is parametrized by user-defined leakage level, which controls maximal inconsistency between initial categorization and resulting clustering. Our method can be implemented as a module in practical expert systems to detect clusters, which combine expert knowledge with true distribution of data. Moreover, it can be used for improving the results of less flexible clustering techniques, such as projection pursuit clustering. The paper presents extensive theoretical analysis of the model and fast algorithm for its efficient optimization. Experimental results show that C3L finds high quality clustering model, which can be applied in discovering meaningful groups in partially classified data. [ABSTRACT FROM AUTHOR]

The use of robots in our daily lives is increasing. As we rely more on robots, thus it becomes more important for us that the robots will continue on with their mission successfully. Unfortunately, these sophisticated, and sometimes very expensive, machines are susceptible to different kinds of faults. It becomes important to apply a Fault Detection (FD) mechanism which is suitable for the domain of robots. Two important requirements of such a mechanism are: high accuracy and low computational-load during operation (online). Supervised learning can potentially produce very accurate FD models, and if the learning takes place offline then the online computational-load can be reduced. Yet, the domain of robots is characterized with the absence of labeled data (e.g., “faulty”, “normal”) required by supervised approaches, and consequently, unsupervised approaches are being used. In this paper we propose a hybrid approach - an unsupervised approach can label a data set, with a low degree of inaccuracy, and then the labeled data set is used offline by a supervised approach to produce an online FD model. Now, we are faced with a choice – should we use the unsupervised or the hybrid fault detector? Seemingly, there is no way to validate the choice due to the absence of (a priori) labeled data. In this paper we give an insight to why, and a tool to predict when, the hybrid approach is more accurate. In particular, the main impacts of our work are (1) we theoretically analyze the conditions under which the hybrid approach is expected to be more accurate. (2) Our theoretical findings are backed with empirical analysis. We use data sets of three different robotic domains: a high fidelity flight simulator, a laboratory robot, and a commercial Unmanned Arial Vehicle (UAV). (3) We analyze how different unsupervised FD approaches are improved by the hybrid technique and (4) how well this improvement fits our prediction tool. The significance of the hybrid approach and the prediction tool is the potential benefit to expert and intelligent systems in which labeled data is absent or expensive to create. [ABSTRACT FROM AUTHOR]

• We propose a proposal for attribute reduction based on stripped quotient sets. • The proposed proposal has solid theoretical background and efficient algorithms. • The proposed proposal reports the superior results when compared to the state-of-the-art methods. Data scenarios on nowadays comprise an enormous number of attributes and instances while not all attributes are necessary and useful for data analytics and knowledge extraction in framing expert and intelligent systems. In such scenarios, removing irrelevant and redundant attributes is a very important task and a challenging research problem. In rough set approach, the attribute reduction method based on the positive region has reported the promising performance in consistent decision tables, while conditional information entropies are employed instead for inconsistent decision tables. However, the complexity of existing methods is still quite high. This paper proposes efficient attribute reduction approaches in inconsistent decision tables by using the concept of stripped quotient sets. Firstly, we present a method for fast determining reducts based on the conditional information entropies including Shannon's and Liang's entropies, called ERED-SQS. Secondly, we define an original notion of attribute reduction based on the stripped quotient set and propose an efficient SRED algorithm to cope with this reduct type. Finally, a simulation performed on 16 datasets varying in a number of objects, up to 5 millions, and a number of attributes, up to 20 thousands, to validate the proposed methods in comparison with the state-of-the-art methods. Experimental results show that the proposed algorithms, ERED-SQS and SRED, have significantly improved the computational time, especially for large data sets. Moreover, the classification accuracy of the reducts obtained from SRED algorithm outperforms that of the reducts based on Shannons and Liangs conditional entropies. This verifies that our proposal is applicable and effective to applications regarding attribute reduction in inconsistent decision tables. [ABSTRACT FROM AUTHOR]

• A distributed predictive model is proposed for Big Data recommendations. • A data partitioning strategy is adopted to speed up Big Data processing. • A novel learning process is devised to accelerate parallel and distributed training. • Matrix factorization and Random forest are used to improve the prediction accuracy. • Our proposal is highly competitive compared with state-of-the-art approaches. Recommendation systems have been widely deployed to address the challenge of overwhelming information. They are used to enable users to find interesting information from a large volume of data. However, in the era of Big Data, as data become larger and more complicated, a recommendation algorithm that runs in a traditional environment cannot be fast and effective. It requires a high computational cost for performing the training task, which may limit its applicability in real-world Big Data applications. In this paper, we propose a novel distributed recommendation solution for Big Data. It is designed based on Apache Spark to handle large-scale data, improve the prediction quality, and address the data sparsity problem. In particular, thanks to a novel learning process, the model is able to significantly speed up the distributed training, as well as improve the performance in the context of Big Data. Experimental results on three real-world data sets demonstrate that our proposal outperforms existing recommendation methods in terms of Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and computational time. [ABSTRACT FROM AUTHOR]

• Provide quick sorting of patients using Alzheimer Disease Intensity Number (ADIN). • Enable estimation of the progression of disease and within a patient's context. • Achieve sufficient expressiveness in understanding patients' disease information. • Knowing the context suggest related treatment(s) or an estimated life expectancy. Modern neuroscience imaging technologies considerably affect diagnostic and prognostic accuracy and facilitate progress towards the cure of brain diseases. The benefits largely depend on the practicalities by which the large-scale imaging and clinical data can be integrated, examined and understood. In EU neuGRID4You (N4U) project, many datasets were generated from research centres and hospitals. In order to perform effective analyses, these datasets and their metadata along with a number of pre-computed parameters are stored in a big data repository. This paper focuses on the patient identification using big data and Fuzzy Logic, which has been achieved through fuzzy processing where a reference number called Alzheimers Disease Identification Number (ADIN) is calculated. It has enabled patients sorting for a particular intensity of Alzheimers disease, short-term estimation of the progression of that disease and context of individual patients with respect to other patients such as appropriate treatment, estimated life expectancy etc. The generated rules define the necessary knowledge base for the inference engine to generate output sets and an aggregate membership function of each rule is formed. Using this function, a most representative value of the total output set is obtained which represents the disease intensity. The implemented system and its evaluation are based on realistic datasets, demonstrators and making use of real-life neuroscience case studies. The presented results of four selected case studies show that this approach have provided sufficient expressiveness in understanding patients' disease information. Finally, a discussion and conclusions are presented on the opportunities offered by the calculation of ADIN to manage Alzheimers disease along with potential future extensions or applications of this work. [ABSTRACT FROM AUTHOR]

• Two phase MapReduce algorithm is proposed for incremental mining of sequential patterns. • Backward mining makes use of the knowledge obtained during the previous mining process. • Co-occurrence reverse map data structure efficiently generates the candidate sequences. • Candidate generation rules avoids the generation of too many false candidates. • Three novel early prune properties are introduced based on the study of item co-occurrences. Sequential Pattern Mining (SPM) is a popular data mining task with broad applications. With the advent of big data, traditional SPM algorithms are not scalable. Hence, many of the researchers have migrated to big data frameworks such as MapReduce and proposed distributed algorithms. However, the existing MapReduce algorithms assume the data as static and do not handle the incremental database updates. Moreover, they use to re-mine the updated database while new sequences are inserted. In this paper, we propose an efficient distributed algorithm for incremental sequential pattern mining (MR-INCSPM) using the MapReduce framework that can handle big data. The proposed algorithm incorporates the backward mining approach that efficiently makes use of the knowledge obtained during the previous mining process. Also, based on the study of item co-occurrences, we propose Co-occurrence Reverse Map (CRMAP) data structure. The issue of combinatorial explosion of candidate sequences is dealt using the proposed CRMAP data structure. Besides, a novel candidate generation and early prune mechanisms are designed using CRMAP to speed up the mining process. The proposed algorithm is evaluated on both the real and synthetic datasets. The experimental results prove the efficacy of MR-INCSPM with respect to processing time, memory and pruning efficiency. [ABSTRACT FROM AUTHOR]

The volatility is related to financial risk and its prediction accuracy is very important in portfolio optimisation. A large body of literature to-date suggests Support Vector Machines (SVM) as the "best of regression" algorithms for financial data regression. Recent work however found that new deep learning––Long Short Term Memory Recurrent Neural Networks (LSTM RNNs) outperformed SVM for classification problems. In the present paper we conduct a new unbiased evaluation of these two modelling techniques for regression problems, and we also compare them with a popular regression model - Generalized Autoregressive Conditional Heteroskedasticity (GARCH) model for financial volatility or risk forecasting. Our experiments using financial data show that the LSTM RNNs performed as good as v -SVR for large interval volatility forecasting and both performed much better than GARCH model for two financial indices (S&P 500 and AAPL). The LSTM RNNS deep learning method can learn from big raw data and can be run with many hidden layers and neurons under GPU to achieve a good prediction for long sequence data compared to the support vector regression. The deep learning technique - LSTM RNNs with big data can be used to improve the volatility prediction instead of v -SVR when the v -SVR does not predict well for some financial stocks of a portfolio. This will help investors to win the competition to maximize their profit. [ABSTRACT FROM AUTHOR]

Crowdsourcing offers the prospect of a more open and socialized approach to product development by bringing together crowdsourcing community resources and connecting open innovation participants. Doing this effectively involves finding an optimal combination of crowdsourcing resources. However, the crowdsourcing community typically focuses on realizing a wide variety of different micro-tasks to meet overall customer requirements, giving rise to numerous potential ways of combining service resources. The parties making up different crowdsourcing community platforms also have various kinds of privacy preferences, making it hard for operators to disclose all of the information relevant to the development of a particular product. This paper proposes a privacy-aware strategy for the optimal combination of crowdsourcing product development resources based on edge computing and cloud computing. Product development application scenarios are examined that envisage the deployment of privacy-aware edge computing and collaboration between a central cloud and edge-based crowdsourcing communities. The crowdsourcing community environment and its privacy risk model are analyzed and the construction and presentation of cloud crowdsourcing platforms for complex tasks are studied. A mechanism for uncovering the optimal cross-platform combination of crowdsourcing product development services based on privacy-aware cloud-edge collaboration is then elaborated, together with a technique for implementing the proposed mechanism via a cooperative bacteria foraging optimization (CBFO) algorithm. Random test data was obtained based on a cloud crowdsourcing platform and two sets of validation tests were undertaken that show that the proposed method can preserve the privacy of crowdsourcing communities, while efficiently solving the problem of the cross-platform mobilization of crowdsourced product development resources. The approach is compared with more basic bacteria foraging optimization (BFO) and genetic algorithm (GA)-based techniques and it is found that CBFO offers better computational performance when confronted with convergence and other factors that can influence the resource combination optimization model, such as a change in the number of micro-tasks. [ABSTRACT FROM AUTHOR]

Choosing the most capable organization to manage biomedical waste (BMW) is a typical multi-attribute group decision-making (MAGDM) problem. MAGDM is frequently used to deal with decision-making scenarios that are fraught with uncertainty and vagueness. As a novel extension of the fuzzy set, the interval-valued Fermatean fuzzy set (IVFFS) can more extensively express the uncertain and vague data in MAGDM issues. In order to combine the information of IVFFS, in this paper, we develop interval-valued Fermatean fuzzy Dombi weighted averaging (geometric) operators with the assistance of Dombi operations. The developed operators can consider the extremely large data sets and the relationships between all decision attributes. Then, utilizing our proposed aggregation operators, we present an integrated MAGDM methodology by combining Preference Ranking Organization METhod for Enrichment Evaluation II (PROMETHEE II) and Stepwise Weight Assessment Ratio Analysis (SWARA) methods. To do this, the attribute weights are estimated by the SWARA method and the PROMETHEE II method determines the preference order of the options. Afterwards, to illustrate the practicality of the proposed methodology, we consider a case study about selecting the most capable organization among the available ones that can handle BMW. The result of this study shows that the bio-chemistry lab is the best BMW management organization. Sensitivity and comparative analysis demonstrate the stability and reliability of the proposed method. According to the findings of this study, we conclude that the proposed methodology offers a comprehensive and systematic approach to evaluating BMW organizations within the IVFF context. [ABSTRACT FROM AUTHOR]

Predictive modeling in medicine involves the development of computational models which are capable of analysing large amounts of data in order to predict healthcare outcomes for individual patients. Computational intelligence approaches are suitable when the data to be modelled are too complex for conventional statistical techniques to process quickly and efficiently. These advanced approaches are based on mathematical models that have been especially developed for dealing with the uncertainty and imprecision which is typically found in clinical and biological datasets. This paper provides a survey of recent work on computational intelligence approaches that have been applied to prostate cancer predictive modeling, and considers the challenges which need to be addressed. In particular, the paper considers a broad definition of computational intelligence which includes metaheuristic optimisation algorithms (also known as nature inspired algorithms), Artificial Neural Networks, Deep Learning, Fuzzy based approaches, and hybrids of these, as well as Bayesian based approaches, and Markov models. Metaheuristic optimisation approaches, such as the Ant Colony Optimisation, Particle Swarm Optimisation, and Artificial Immune Network have been utilised for optimising the performance of prostate cancer predictive models, and the suitability of these approaches are discussed. [ABSTRACT FROM AUTHOR]

Real-world optimization problems typically involve multiple objectives to be optimized simultaneously under multiple constraints and with respect to several variables. While multi-objective optimization itself can be a challenging task, equally difficult is the ability to make sense of the obtained solutions. In this two-part paper, we deal with data mining methods that can be applied to extract knowledge about multi-objective optimization problems from the solutions generated during optimization. This knowledge is expected to provide deeper insights about the problem to the decision maker, in addition to assisting the optimization process in future design iterations through an expert system. The current paper surveys several existing data mining methods and classifies them by methodology and type of knowledge discovered. Most of these methods come from the domain of exploratory data analysis and can be applied to any multivariate data. We specifically look at methods that can generate explicit knowledge in a machine-usable form. A framework for knowledge-driven optimization is proposed, which involves both online and offline elements of knowledge discovery. One of the conclusions of this survey is that while there are a number of data mining methods that can deal with data involving continuous variables, only a few ad hoc methods exist that can provide explicit knowledge when the variables involved are of a discrete nature. Part B of this paper proposes new techniques that can be used with such datasets and applies them to discrete variable multi-objective problems related to production systems. [ABSTRACT FROM AUTHOR]

In this paper, we introduce a new adaptive rule-based classifier for multi-class classification of biological data, where several problems of classifying biological data are addressed: overfitting, noisy instances and class-imbalance data. It is well known that rules are interesting way for representing data in a human interpretable way. The proposed rule-based classifier combines the random subspace and boosting approaches with ensemble of decision trees to construct a set of classification rules without involving global optimisation. The classifier considers random subspace approach to avoid overfitting, boosting approach for classifying noisy instances and ensemble of decision trees to deal with class-imbalance problem. The classifier uses two popular classification techniques: decision tree and k-nearest-neighbor algorithms. Decision trees are used for evolving classification rules from the training data, while k-nearest-neighbor is used for analysing the misclassified instances and removing vagueness between the contradictory rules. It considers a series of k iterations to develop a set of classification rules from the training data and pays more attention to the misclassified instances in the next iteration by giving it a boosting flavour. This paper particularly focuses to come up with an optimal ensemble classifier that will help for improving the prediction accuracy of DNA variant identification and classification task. The performance of proposed classifier is tested with compared to well-approved existing machine learning and data mining algorithms on genomic data (148 Exome data sets) of Brugada syndrome and 10 real benchmark life sciences data sets from the UCI (University of California, Irvine) machine learning repository. The experimental results indicate that the proposed classifier has exemplary classification accuracy on different types of biological data. Overall, the proposed classifier offers good prediction accuracy to new DNA variants classification where noisy and misclassified variants are optimised to increase test performance. [ABSTRACT FROM AUTHOR]

Highlights • A novel SVM is proposed for high noise and unbalanced distribution data. • Feature selection is improved by a novel method using the hierarchical relations in feature sets. • Roulette algorithm is introduced into cooperative coevolution. Abstract In real applications, noises are often present in the obtained data, which would considerably affect the performance of machine learning models. Although support vector machine (SVM) is a classic and efficient learning model, however, it is sensitive to noises in the training data. In this paper, a novel support vector machine named as reconstructed training-set SVM (RTS-SVM) is proposed to implement classification for high-noise data, where the roulette cooperative coevolution algorithm (R-CC) is used to optimize the parameters of RTS-SVM. The proposed SVM model is applicable to make the classification of high-noise data by tackling with the sensitive effect of the ''soft margin'' of SVM on the original training set. By means of the hierarchical relations existing in feature sets, hierarchical grouping (HG) algorithm is applied to construct feature subsets, based on which R-CC coordinates the parameters of RTS-SVM to achieve the optimization of the whole model. The application of the proposed scheme in the classification of financial time series is mainly discussed. Besides, the proposed model is also verified by using synthetic data with high noises and daily life data sets. Examples are provided to illustrate the effectiveness and practicability of the proposed algorithm. [ABSTRACT FROM AUTHOR]

Highlights • Enabling Radial Basis Function for surface fitting using Big Data and limited memory. • Solution for missing value prediction on Big Data. • Performance for missing value prediction is found better than Kernel Ridge Regression. • Establishes bias variance tradeoff with the help of hyperparameters. Abstract The problem of missing values in the process of data acquisition is becoming critical due to both hardware failure and human error. Radial Basis Function based interpolation and missing value prediction in the dependent variable through surface fitting has been the viable solution from a long time. However, the solution works well on the idea of building one equation of 'N' weight variables corresponding to each sampled data point from a set of 'N' samples. Still, because of the inherent computations, it demands big primary memory when 'N' becomes large. Hence, on a memory-restricted setup, sometimes it suffers from the memory overflow problem. In this paper, we propose a novel data decomposition based RBF enabled surface fitting approach by building and aggregating multiple small models in an overlapping manner, which works well even with smaller primary memory with minimal impact on loss of generality. We also consider two hyperparameters as window size and overlapping index in order to tune bias-variance tradeoff. The proposed approach is applied to ten real-world datasets having multiple dimensions and results are found competitive while comparing with single trained model and Kernel Ridge Regression. Therefore, we believe this approach will rejuvenate the RBF based surface fitting method in attaining better performance in the Big data world too. [ABSTRACT FROM AUTHOR]

Highlights • Two schemes are proposed on distrust propagation in influence maximization. • Three information diffusion models are proposed considering trust and distrust. • Influence spread is modeled more accurately applying distrust relationships. • When a distrusted user performs an action, the target users may tend not to do it. Abstract Influence maximization is an important issue in social network analysis domain which concerns finding the most influential nodes. Determining the influential nodes is made with respect to information diffusion models. Most of the existing models only contain trust relationships while distrust exist in social networks as well. There exist some drawbacks in limited studies where distrust relationship is involved. The most outstanding drawback is the lack of assessment on the validity of the schemes presented on how influence propagates through distrust relationships in comparison with real word propagation in social networks. In this paper, two schemes are proposed, where based on each, some new models are proposed in two classes: cascade-based and threshold-based. All models of concern here are evaluated in comparison with the benchmark models through two real data sets, the Epinions and Bitcoin OTC. Results obtained indicate the superiority of one of the proposed schemes: when a distrusted user performs an action or adopts an opinion, the target users may tend not to do it. [ABSTRACT FROM AUTHOR]

Highlights • A semantic approach to represent and validate Big Data analytics is proposed. • An OWL Ontology and SWRL rules are developed for reasoning in workflow design. • The proposal is validated with two real-world (traffic) and academic cases study. • Obtained semantized data successfully recommends and validate Big Data tasks. • We provide actual Big Data practitioners with software to enhance their analytics. Abstract Knowledge extraction and incorporation is currently considered to be beneficial for efficient Big Data analytics. Knowledge can take part in workflow design, constraint definition, parameter selection and configuration, human interactive and decision-making strategies. This paper proposes BIGOWL, an ontology to support knowledge management in Big Data analytics. BIGOWL is designed to cover a wide vocabulary of terms concerning Big Data analytics workflows, including their components and how they are connected, from data sources to the analytics visualization. It also takes into consideration aspects such as parameters, restrictions and formats. This ontology defines not only the taxonomic relationships between the different concepts, but also instances representing specific individuals to guide the users in the design of Big Data analytics workflows. For testing purposes, two case studies are developed, which consists in: first, real-world streaming processing with Spark of traffic Open Data, for route optimization in urban environment of New York city; and second, data mining classification of an academic dataset on local/cloud platforms. The analytics workflows resulting from the BIGOWL semantic model are validated and successfully evaluated. [ABSTRACT FROM AUTHOR]

Feature selection is an important preprocessing task in classification that eliminates the irrelevant, redundant, and noisy features. Improving the performance of model, decreasing the computational cost, and adjusting the “curse of dimensionality” are the key advantages of feature selection task. The evolution process of the existing multi-objective based feature selection algorithms is relied on the objective space while the problem space contains useful information. This paper proposes a multi-objective PSO based method named RFPSOFS that ranks the features based on their frequencies in the archive set. Then, these ranks are used to refine the archive set and guide the particles. The proposed method is compared with three PSO based and one genetic based multi-objective methods on 9 Benchmark datasets. Qualitative and quantitative analyses of the results are performed by visual analysis of the Pareto fronts and three performance metrics respectively. Finally, remarkable performance in datasets with more than hundred features and satisfactory performance in datasets with less than hundred features are obtained. [ABSTRACT FROM AUTHOR]

With the ubiquity of the Internet, platforms such as: Google, Wikipedia and the like can provide insights pertaining to firms’ financial performance as well as capture the collective interest of traders through search trends, number of web page visitors and/or financial news sentiment. Information emanating from these platforms can significantly affect, or be affected by, changes in the stock market. The overarching goal of this paper is to develop a financial expert system that incorporates these features to predict short term stock prices. Our expert system is comprised of two main modules: a knowledge base and an artificial intelligence (AI) platform. The “knowledge base” for our expert system captures: (a) historical stock prices; (b) several well-known technical indicators; (c) counts and sentiment scores of published news articles for a given stock; (d) trends in Google searches for the given stock ticker; and (e) number of unique visitors for pertinent Wikipedia pages. Once the data is collected, we use a structured approach for data preparation. Then, the AI platform trains four machine learning ensemble methods: (a) a neural network regression ensemble; (b) a support vector regression ensemble; (c) a boosted regression tree; and (d) a random forest regression. In the cross-validation phase, the AI platform picks the “best” ensemble for a given stock. To evaluate the efficacy of our expert system, we first present a case study based on the Citi Group stock ($C) with data collected from 01/01/2013 - 12/31/2016. We show the expert system can predict the 1-day ahead $C stock price with a mean absolute percent error (MAPE)  ≤  1.50% and the 1–10 day ahead with a MAPE  ≤  1.89%, which is better than the reported results in the literature. We show that the use of features extracted from online sources does not substitute the traditional financial metrics, but rather supplements them to improve upon the prediction performance of machine learning based methods. To highlight the utility and generalizability of our expert system, we predict the 1-day ahead price of 19 additional stocks from different industries, volatilities and growth patterns. We report an overall mean for the MAPE statistic of 1.07% across our five different machine learning models, including a MAPE of under 0.75% for 18 of the 19 stocks for the best ensemble (boosted regression tree). [ABSTRACT FROM AUTHOR]

In this paper, a formal method is presented to: (1) establish a system to visualize huge quantities of related information (2) graphically show the degree of similarity between these related documents (3) navigate between related documents so that the views are reconfigured in a dynamic and gradual way. This method uses, as input, a set of documents that have been related using any existing algorithm (relationships based on contents, on votes to contents, on social relations, on demographic information, etc.). The proposed method includes two different visualization and navigational versions related to: (a) visualization by explicit graph and (b) visualization by implicit graph. As a summary, the following characteristics are presented: (1) this method maximizes the information that is displayed at each moment, so that a balance is found between: (a) the number of relations shown, (b) the number of documents viewed, (c) the amount of information presented by each document and (d) how to show the degree of similarity between documents, (2) the method includes specific rules of interaction with the user to allow navigation in a simple, gradual and intuitive way. [ABSTRACT FROM AUTHOR]

Digital journalism has faced a dramatic change and media companies are challenged to use data science algorithms to be more competitive in a Big Data era. While this is a relatively new area of study in the media landscape, the use of machine learning and artificial intelligence has increased substantially over the last few years. In particular, the adoption of data science models for personalization and recommendation has attracted the attention of several media publishers. Following this trend, this paper presents a research literature analysis on the role of Data Science (DS) in Digital Journalism (DJ). Specifically, the aim is to present a critical literature review, synthetizing the main application areas of DS in DJ, highlighting research gaps, challenges, and opportunities for future studies. Through a systematic literature review integrating bibliometric search, text mining, and qualitative discussion, the relevant literature was identified and extensively analyzed. The review reveals an increasing use of DS methods in DJ, with almost 47% of the research being published in the last three years. An hierarchical clustering highlighted six main research domains focused on text mining, event extraction, online comment analysis, recommendation systems, automated journalism, and exploratory data analysis along with some machine learning approaches. Future research directions comprise developing models to improve personalization and engagement features, exploring recommendation algorithms, testing new automated journalism solutions, and improving paywall mechanisms. [ABSTRACT FROM AUTHOR]

With the increasing popularity of large scale Knowledge Graph (KG)s, many applications such as semantic analysis, search and question answering need to link entity mentions in texts to entities in KGs. Because of the polysemy problem in natural language, entity disambiguation is thus a key problem in current research. Existing disambiguation methods have considered entity prominence, context similarity and entity-entity relatedness to discriminate ambiguous entities, which are mainly working on document or paragraph level texts containing rich contextual information, and based on lexical matching for computing context similarity. When meeting short texts containing limited contextual information, such as web queries, questions and tweets, those conventional disambiguation methods are not good at handling single entity mention and measuring context similarity. In order to enhance the performance of disambiguation methods based on context similarity with such short texts, we propose SCSNED method for disambiguation based on semantic similarity between contextual words and informative words of entities in KGs. Specially, we exploit the effectiveness of both knowledge-based and corpus-based semantic similarity methods for entity disambiguation with SCSNED. Moreover, we propose a Category2Vec embedding model based on joint learning of word and category embedding, in order to compute word-category similarity for entity disambiguation. We show the effectiveness of these proposed methods with illustrative examples, and evaluate their effectiveness in a comparative experiment for entity disambiguation in real world web queries, questions and tweets. The experimental results have identified the effectiveness of different semantic similarity methods, and demonstrated the improvement of semantic similarity methods in SCSNED and Category2Vec over the conventional context similarity baseline. We further compare the proposed approaches with the state of the art entity disambiguation systems and show the performances of the proposed approaches are among the best performing systems. In addition, one important feature of the proposed approaches using semantic similarity, is the potential application on any existing KGs since they mainly use common features of entity descriptions and categories. Another contribution of the paper is an updated survey on background of entity disambiguation in KGs and semantic similarity methods. [ABSTRACT FROM AUTHOR]

Semi-supervised clustering generally assumes that prior knowledge is helpful to improve clustering performance. However, the prior knowledge may degenerate the clustering performance if one collects wrong information, such as wrong labels. Hence, it is meaningful to design a safe semi-supervised clustering method which never performs worse than the corresponding unsupervised and semi-supervised clustering methods. In this paper, we develop local homogeneous consistent safe semi-supervised clustering where class labels are given as the prior knowledge. To the best of our knowledge, it is the first time safe semi-supervised clustering has been studied. The basic idea is that the predictions of a labeled sample and its nearest homogeneous unlabeled ones should be similar when the labeled one is risky. In our algorithm, we firstly build a local graph to model the relationship between the labeled sample and its nearest homogeneous unlabeled ones through the results obtained by unsupervised clustering. A graph-based regularization term is then constructed to allow the predictions of the labeled samples to approach that of the local homogeneous neighbors. It is expected to reduce the risk of the labeled samples. Meanwhile, our algorithm positively exploits the labeled samples by restricting the corresponding outputs to be the given class labels when the labeled ones may be helpful. In this sense, the predictions of the labeled samples in our algorithm are a tradeoff between the given class labels and the predictions of local homogeneous neighbors. To verify the effectiveness of our algorithm, we conduct a series of experiments on several UCI datasets. The results show that our algorithm outperforms the corresponding unsupervised and semi-supervised clustering methods even if the wrongly labeled ratio reaches 30%. In this sense, the proposed algorithm will not only enrich the theoretical knowledge in the machine learning field, but significantly improve the practicability of semi-supervised clustering in the expert and intelligent systems. [ABSTRACT FROM AUTHOR]

In the era of big data and networking, it is necessary to develop a secure and robust digital watermarking scheme with high computational efficiency to protect copyrights of digital works. However, most of the existing methods focus on robustness and embedding capacity, losing sight of security or requiring significant computational resources in the encryption process. This paper proposed a new digital image watermarking model based on scrambling algorithm Logistic and RSA asymmetric encryption algorithm to guarantee the security of the hidden data at the foundation of large embedding capacity, good robustness and high computational efficiency. The experiments involved applying the encryption algorithms of Logistic and RSA to the watermark image and performing the hybrid decomposition of Discrete wavelet transform (DWT) and Singular Value Decomposition (SVD) on the host image, and the watermark was embedded into the low-frequency sub-band of the host. The values of PSNR and NCC were measured to estimate the imperceptibility and robustness of the proposed watermarking scheme, and the CPU running time was recorded to measure the complexity of the proposed main algorithm in execution time. Experimental results showed the superiority of the proposed watermarking scheme. [ABSTRACT FROM AUTHOR]

In this paper we present PFDCMSS (Parallel Forward Decay Count–Min Space Saving) which, to the best of our knowledge, is the world first message–passing parallel algorithm for mining time–faded heavy hitters. The algorithm is a parallel version of the recently published FDCMSS (Forward Decay Count–Min Space Saving) sequential algorithm. We formally prove its correctness by showing that the underlying data structure, a sketch augmented with a Space Saving stream summary holding exactly two counters, is mergeable. Whilst mergeability of traditional sketches derives immediately from theory, we show that, instead, merging our augmented sketch is non trivial. Nonetheless, the resulting parallel algorithm is fast and simple to implement. The very large volumes of modern datasets in the context of Big Data present new challenges that current sequential algorithms can not cope with; on the contrary, parallel computing enables near real time processing of very large datasets, which are growing at an unprecedented scale. Our algorithm’s implementation, taking advantage of the MPI (Message Passing Interface) library, is portable, reliable and provides cutting–edge performance. Extensive experimental results confirm that PFDCMSS retains the extreme accuracy and error bound provided by FDCMSS whilst providing excellent parallel scalability. Our contributions are three-fold: (i) we prove the non trivial mergeability of the augmented sketch used in the FDCMSS algorithm; (ii) we derive PFDCMSS, a novel message–passing parallel algorithm; (iii) we experimentally prove that PFDCMSS is extremely accurate and scalable, allowing near real time processing of large datasets. The result supports both casual users and seasoned, professional scientists working on expert and intelligent systems. [ABSTRACT FROM AUTHOR]

This paper learns the route choice behavior of passengers from Auto Fare Collection, timetable, and train loading data using a method combined with Bayesian inference and Metropolis-Hasting sampling. First, the influential factors of route choice such as in-vehicle travel time, transfer time, and in-vehicle crowding are given. Next, formulations are established based on AFC, timetable and train loading data, which are merged into a logit model of route choice behavior of subway passengers. Next, an algorithm integrating Bayesian inference and Metropolis-Hasting sampling is designed to calibrate parameters of the logit model. Finally, a case study of Beijing subway is applied to verify the validity of the model and algorithm. A detailed discussion shows that in-vehicle crowding plays a crucial role in passenger route choice behavior. [ABSTRACT FROM AUTHOR]

This paper explores advances in the data mining field to solve the fundamental Document Information Retrieval problem. In the proposed approach, useful knowledge is first discovered by using data mining techniques, then swarms use this knowledge to explore the whole space of documents intelligently. We have investigated two data mining techniques in the preprocessing step. The first one aims to split the collection of documents into similar clusters by using the K-means algorithm, while the second one extracts the most closed frequent terms on each cluster already created using the DCI_Closed algorithm. For the solving step, BSO (Bees Swarm Optimization) is used to explore the cluster of documents deeply. The proposed approach has been evaluated on well-known collections such as CACM (Collection of ACM), TREC (Text REtrieval Conference), Webdocs, and Wikilinks, and it has been compared to state-of-the-art data mining, bio-inspired and other documents information retrieval based approaches. The results show that the proposed approach improves the quality of returned documents considerably, with a competitive computational time compared to state-of-the-art approaches. [ABSTRACT FROM AUTHOR]

E -communication represents a major threat to users who are exposed to a number of risks and potential attacks. Detecting these risks with as much anticipation as possible is crucial for prevention. However, much research so far has focused on forensic tools that can be applied only when an attack has been performed. This paper proposes a novel and effective methodology for the early detection of threats in written social media. The goal is to recognize a potential attack before it is consummated, and using a minimum amount of information. The proposed approach considers the use of profile-based representations (PBRs) for this goal. PBRs have multiple benefits, including non-sparsity, low dimensionality, and a proved discriminative power. Moreover, representations for partial documents can be derived naturally with PBRs, which makes them suitable for the addressed problem. Results include empirical evidence on the usefulness of PBRs in the early recognition setting for two tasks in which anticipation is critical: sexual predator detection and aggressive text identification. These results reveal, on the one hand, that PBRs achieve state of the art performance when using full-length documents (i.e., the classical task), and, on the other hand, that the proposed methodology outperforms previous work on early recognition of sexual predators by a considerable margin, while obtaining state of the art performance in aggressive text identification. To the best of our knowledge, these are the best results reported on early recognition for the approached problems. We foresee this work will pave the way for the development of novel methodologies for the problem and will motivate further research from the intelligent systems and text mining communities. [ABSTRACT FROM AUTHOR]

Feature selection, both for supervised as well as for unsupervised classification is a relevant problem pursued by researchers for decades. There are multiple benchmark algorithms based on filter, wrapper and hybrid methods. These algorithms adopt different techniques which vary from traditional search-based techniques to more advanced nature inspired algorithm based techniques. In this paper, a hybrid feature selection algorithm using graph-based technique has been proposed. The proposed algorithm has used the concept of Feature Association Map (FAM) as an underlying foundation. It has used graph-theoretic principles of minimal vertex cover and maximal independent set to derive feature subset. This algorithm applies to both supervised and unsupervised classification. The performance of the proposed algorithm has been compared with several benchmark supervised and unsupervised feature selection algorithms and found to be better than them. Also, the proposed algorithm is less computationally expensive and hence has taken less execution time for the publicly available datasets used in the experiments, which include high-dimensional datasets. [ABSTRACT FROM AUTHOR]

Feature selection is a crucial machine learning technique aimed at reducing the dimensionality of the input space. By discarding useless or redundant variables, not only it improves model performance but also facilitates its interpretability. The well-known Support Vector Machines–Recursive Feature Elimination (SVM-RFE) algorithm provides good performance with moderate computational efforts, in particular for wide datasets. When using SVM-RFE on a multiclass classification problem, the usual strategy is to decompose it into a series of binary ones, and to generate an importance statistics for each feature on each binary problem. These importances are then averaged over the set of binary problems to synthesize a single value for feature ranking. In some cases, however, this procedure can lead to poor selection. In this paper we discuss six new strategies, based on list combination, designed to yield improved selections starting from the importances given by the binary problems. We evaluate them on artificial and real-world datasets, using both One–Vs–One (OVO) and One–Vs–All (OVA) strategies. Our results suggest that the OVO decomposition is most effective for feature selection on multiclass problems. We also find that in most situations the new K-First strategy can find better subsets of features than the traditional weight average approach. [ABSTRACT FROM AUTHOR]

Computer-aided diagnosis (CAD) systems have drawn attention of researchers for arriving at qualitative and faster clinical decisions, and hence has become one of the most important directions of research. In this paper, we propose an efficient CAD system to classify pathological and healthy brains using brain MR images. The suggested pathological brain detection system (PBDS) has the ability to help radiologists to initiate the corrective measures for treating the ailing patients at an early stage. The proposed scheme uses a simplified pulse-coupled neural network (SPCNN) for the region of interest (ROI) segmentation and fast discrete curvelet transform (FDCT) for feature extraction. Subsequently, PCA+LDA approach is harnessed for feature dimensionality reduction and finally probabilistic neural network (PNN) is applied for classification. The scheme is validated on various standard datasets and compared with existing competent schemes with respect to classification accuracy and number of features. The statistical set up is kept similar as reported in the recent literature to derive an unbiased analysis. Experimental results demonstrate that the suggested scheme yields higher accuracy as compared to others with considerably less number of features. The number of parameters need to be tuned at different stages are significantly less in contrast to existing schemes. Further, PNN used has a simple network structure and offers faster learning speed. Therefore, the proposed scheme can effectively detect pathological brain in real-time and hence has a potential to be installed on medical robots. [ABSTRACT FROM AUTHOR]

This paper presents an improved version of the well-established k nearest neighbor ( k -NN) and fuzzy NN (FNN), termed the multi-objective genetic-algorithm-modified FNN (MOGA-MFNN). The MFNN design problem is converted into a multi-modal objective maximization problem constrained by four objective functions. Thereafter, the associated parameter set of the MFNN and the feature attributes can be determined optimally and automatically via the non-dominated sorting genetic algorithm II. We introduce two new objective functions termed the Margin-I and Margin-II, which are used to improve the generalization capability of the MFNN for the unknown data, along with two existing performance functions: the geometric mean and the area under the receiver-operated characteristic curve for the training accuracy. Moreover, we proposed a novel data-dependent weight-assignment technique for local class membership functions of the MFNN. The technique enables the MFNN to determine its local neighbors adaptively through the MOGA algorithm. To expedite the classification, the MOGA-MFNN is implemented on a graphical processing unit (GPU), which significantly increases the computation speed. Furthermore, the local class-membership function of the MFNN can be computed in advance, rather than delaying it to the classification stage. This again can improve the classification speed. The MOGA-MFNN is evaluated on 20 datasets obtained from the repository of the University of California, Irvine (UCI). The experiments with rigorous statistical significance tests demonstrate that the proposed method performs competitively with the existing methods. [ABSTRACT FROM AUTHOR]

Due to its complex structure, production form a shale-gas formation requires more drillings than those for the traditional reservoirs. Modeling of such reservoirs and making predictions for their production also require highly extensive datasets. Both are very costly. In-situ measurements, such as well-logging, are one of most indispensable tools for providing considerable amount of information and data for such unconventional reservoirs. Production from shale reservoirs involves the so-called fracking, i.e. injection of water and chemicals into the formation in order to open up flow paths for the hydrocarbons. The measurements and any other types of data are utilized for making critical decisions regarding development of a potential shale reservoir, as it requires hundreds of millions of dollar initial investment. The questions that must be addressed include, does the region under study can be used economically for producing hydrocarbons? If the response to the first question is affirmative, then, where are the best places to carry out hydro-fracking? Through the answers to such questions one identifies the sweet spots of shale reservoirs, which are the regions that contain high total organic carbon (TOC) and brittle rocks that can be fractured. In this paper, two methods from data mining and machine learning techniques are used to aid identifying such regions. The first method is based on a stepwise algorithm that determines the best combination of the variables (well-log data) to predict the target parameters. However, in order to incorporate more training, and efficiently use the available datasets, a hybrid machine-learning algorithm is also presented that models more accurately the complex spatial correlations between the input and target parameters. Then, statistical comparisons between the estimated variables and the available data are made, which indicate very good agreement between the two. The proposed model can be used effectively to estimate the probability of targeting the sweet spots. In the light of an automatic input and parameter selection, the algorithm does not require any further adjustment and can continuously evaluate the target parameters, as more data become available. Furthermore, the method is able to optimally identify the necessary logs that must be run, which significantly reduces data acquisition operations. [ABSTRACT FROM AUTHOR]

Fault prediction is an important topic for the industry as, by providing effective methods for predictive maintenance, allows companies to perform important time and cost savings. In this paper we describe an application developed to predict and explain door failures on metro trains. To this end, the aim was twofold: first, devising prediction techniques capable of early detecting door failures from diagnostic data; second, describing failures in terms of properties distinguishing them from normal behavior. Data pre-processing was a complex task aimed at overcoming a number of issues with the dataset, like size, sparsity, bias, burst effect and trust. Since failure premonitory signals did not share common patterns, but were only characterized as non-normal device signals, fault prediction was performed by using outlier detection. Fault explanation was finally achieved by exhibiting device features showing abnormal values. An experimental evaluation was performed to assess the quality of the proposed approach. Results show that high-degree outliers are effective indicators of incipient failures. Also, explanation in terms of abnormal feature values (responsible for outlierness) seems to be quite expressive.There are some aspects in the proposed approach that deserve particular attention. We introduce a general framework for the failure detection problem based on an abstract model of diagnostic data, along with a formal problem statement. They both provide the basis for the definition of an effective data pre-processing technique where the behavior of a device, in a given time frame, is summarized through a number of suitable statistics. This approach strongly mitigates the issues related to data errors/noise, thus enabling to perform an effective outlier detection. All this, in our view, provides the grounds of a general methodology for advanced prognostic systems. [ABSTRACT FROM AUTHOR]

Active learning is used for classification when labeling data are costly, while the main challenge is to identify the critical instances that should be labeled. Clustering-based approaches take advantage of the structure of the data to select representative instances. In this paper, we developed the active learning through density peak clustering (ALEC) algorithm with three new features. First, a master tree was built to express the relationships among the nodes and assist the growth of the cluster tree. Second, a deterministic instance selection strategy was designed using a new importance measure. Third, tri-partitioning was employed to determine the action to be taken on each instance during iterative clustering, labeling, and classifying. Experiments were performed with 14 datasets to compare against state-of-the-art active learning algorithms. Results demonstrated that the new algorithm had higher classification accuracy using the same number of labeled data. [ABSTRACT FROM AUTHOR]

In the era of big data, some data records are interrelated with each other in many areas, such as marketing, management, health care, and education. These interrelated data can be more naturally represented as networks with nodes and edges. Inside this type of networks, there is usually a hidden community structure where each community represents a relatively independent functional module. Such hidden community structures are useful for many applications, such as word-of-mouth marketing, promoting decentralized social interactions inside organizations, and searching biological pathways related to various diseases. Therefore, how to detect hidden community structures becomes an important task with wide applications. Currently, modularity-based methods are widely-used among many existing community structure detection methods. They detect communities with more internal edges than expected under the null hypothesis of independence. Since research in correlation analysis also searches for patterns which occur more than expected under the null hypothesis of independence, this paper proposed a framework of changing the original modularity function according to different existing correlation functions in the correlation analysis research area. Such a framework can utilize not only the current but also the future potential research progresses in correlation analysis to advance community detection. In addition, a novel graphical analysis on different modified-modularity functions is conducted to analyze their different preferences, which are also validated by our evaluation on both real life and simulated networks. Our work to connect modularity-based methods with correlation analysis has several significant impacts on the community detection research and its applications to expert and intelligent systems. First, the research progress in correlation analysis can be utilized to define a more effective objective function in community detection for better detection results since different real-life applications might need communities with different resolutions. Second, any existing research progress for the modularity function, such as the Louvain method for speeding up the search and different extensions for overlapping community detection, can be applied in a similar way to the new objective function derived from existing correlation functions, because the new objective function is unified within one framework with the modularity function. Third, our framework opens a large unexplored area for the researchers interested in community detection. For example, what is the best heuristic search method for each different objective function? What are the characteristics of each objective function when applied to overlapping community detection? Among different extensions to overlapping community detection, which extension is better for each objective function? [ABSTRACT FROM AUTHOR]

Most of existing text automatic summarization algorithms are targeted for multi-documents of relatively short length, thus difficult to be applied immediately to novel documents of structure freedom and long length. In this paper, aiming at novel documents, we propose a topic modeling based approach to extractive automatic summarization, so as to achieve a good balance among compression ratio, summarization quality and machine readability. First, based on topic modeling, we extract the candidate sentences associated with topic words from a preprocessed novel document. Second, with the goals of compression ratio and topic diversity, we design an importance evaluation function to select the most important sentences from the candidate sentences and thus generate an initial novel summary. Finally, we smooth the initial summary to overcome the semantic confusion caused by ambiguous or synonymous words, so as to improve the summary readability. We evaluate experimentally our proposed approach on a real novel dataset. The experiment results show that compared to those from other candidate algorithms, each automatic summary generated by our approach has not only a higher compression ratio, but also better summarization quality. [ABSTRACT FROM AUTHOR]

In this paper, we propose a novel Evolutionary Mamdani Fuzzy Rule-based System named DECO 3 RUM ( D ifferential E volution based Co operative and Co mpeting learning of Co mpact Ru le-based M odels). The main advantage of DECO 3 RUM is that it offers simple and descriptive Mamdani models, without sacrificing the accuracy. It has further been designed to work with Big Data problems, where the number of patterns or features is significant. This is achieved through the successive use of three stages: (a) an instance selection stage, (b) an evolutionary rule base learning stage following the Genetic Cooperative Competitive Learning approach (where the novel Fuzzy Token Competition method based on AdaBoost tackles the cooperation-competition problem), which utilizes Differential Evolution as its learning algorithm, and (c) a post-processing stage to simplify the derived model and enhance its accuracy. DECO 3 RUM was validated using 24 datasets, and was additionally applied to one real-world high dimensional dataset from the domain of soil science. Through the comparisons with other state of the art evolutionary fuzzy systems it was demonstrated that DECO 3 RUM achieves a fair balance between accuracy and interpretability, performing better than its Mamdani counterparts. [ABSTRACT FROM AUTHOR]

Learning from imbalanced datasets is challenging for standard algorithms, as they are designed to work with balanced class distributions. Although there are different strategies to tackle this problem, methods that address the problem through the generation of artificial data constitute a more general approach compared to algorithmic modifications. Specifically, they generate artificial data that can be used by any algorithm, not constraining the options of the user. In this paper, we present a new oversampling method, Self-Organizing Map-based Oversampling (SOMO), which through the application of a Self-Organizing Map produces a two dimensional representation of the input space, allowing for an effective generation of artificial data points. SOMO comprises three major stages: Initially a Self-Organizing Map produces a two-dimensional representation of the original, usually high-dimensional, space. Next it generates within-cluster synthetic samples and finally it generates between cluster synthetic samples. Additionally we present empirical results that show the improvement in the performance of algorithms, when artificial data generated by SOMO are used, and also show that our method outperforms various oversampling methods. [ABSTRACT FROM AUTHOR]

F-score is a simple feature selection technique, however, it works only for two classes. This paper proposes a novel feature ranking method based on Fisher discriminate analysis (FDA) and F-score, denoted as FDAF-score, which considers the relative distribution of classes in a multi-dimensional feature space. The main idea is that a proper subset is got according to maximizing the proportion of average between-class distance to the relative within-class scatter. Because the method removes all insignificant features at a time, it can effectively reduce computational cost. Experiments on six benchmarking UCI datasets and two artificial datasets demonstrate that the proposed FDAF-score algorithm can not only obtain good results with fewer features than the original datasets as well as fast computation but also deal with the classification problem with noises well. [ABSTRACT FROM AUTHOR]

The entity reconciliation (ER) problem aroused much interest as a research topic in today's Big Data era, full of big and open heterogeneous data sources. This problem poses when relevant information on a topic needs to be obtained using methods based on: (i) identifying records that represent the same real world entity, and (ii) identifying those records that are similar but do not correspond to the same real-world entity. ER is an operational intelligence process, whereby organizations can unify different and heterogeneous data sources in order to relate possible matches of non-obvious entities. Besides, the complexity that the heterogeneity of data sources involves, the large number of records and differences among languages, for instance, must be added. This paper describes a Systematic Mapping Study (SMS) of journal articles, conferences and workshops published from 2010 to 2017 to solve the problem described before, first trying to understand the state-of-the-art, and then identifying any gaps in current research. Eleven digital libraries were analyzed following a systematic, semiautomatic and rigorous process that has resulted in 61 primary studies. They represent a great variety of intelligent proposals that aim to solve ER. The conclusion obtained is that most of the research is based on the operational phase as opposed to the design phase, and most studies have been tested on real-world data sources, where a lot of them are heterogeneous, but just a few apply to industry. There is a clear trend in research techniques based on clustering/blocking and graphs, although the level of automation of the proposals is hardly ever mentioned in the research work. [ABSTRACT FROM AUTHOR]

The arrival of new technologies related to smart grids and the resulting ecosystem of applications and management systems pose many new problems. The databases of the traditional grid and the various initiatives related to new technologies have given rise to many different management systems with several formats and different architectures. A heterogeneous data source integration system is necessary to update these systems for the new smart grid reality. Additionally, it is necessary to take advantage of the information smart grids provide. In this paper, the authors propose a heterogeneous data source integration based on IEC standards and metadata mining. Additionally, an automatic data mining framework is applied to model the integrated information. [ABSTRACT FROM AUTHOR]