Search

With the rise of ready-to-assemble furniture, driven by international giants like IKEA, assembly services were increasingly offered by the same retailers. When planning orders with assembly services, the estimation of the service time leads to additional difficulties compared to standard delivery planning. Assembling large wardrobes or kitchens can take hours or even days while assembling a chair can be done in a few minutes. Combined with the usually vast amounts of offered products, a lot of knowledge is required to plan efficient and exact delivery routes. This paper shows how an artificial neural network (ANN) can be used to accurately predict the service time of a delivery based on factors such as the goods to be delivered or the personnel providing the service. The data used include not only deliveries with assembly of furniture, but also deliveries of goods without assembly and delivery of goods requiring electrical installation. The goal is to create a solution that can predict the time needed based on criteria such the type of furniture, the weight of the goods, and the experiences of the service technicians. The findings show that ANNs can be applied to this scenario and outperform more classical approaches, such as multiple linear regression or support vector machines. Still existing problems are largely due to the provided data, e.g., a large difference between the number of short and longer duration orders, which made it harder to accurately predict orders with longer duration. [ABSTRACT FROM AUTHOR]

Accurate stock price prediction is critical for investment decisions in the stock market. To improve the performance of stock price prediction, this paper proposes a novel two-stage prediction model that consists of a decomposition algorithm, a nonlinear ensemble strategy, and three individual machine learning models. Specifically, in the first stage, the stock price time series is decomposed into a finite number of sub-series by variational mode decomposition (VMD). Subsequently, three individual machine learning models, namely support vector machine regression (SVR), extreme learning machine (ELM), and deep neural network (DNN), are separately employed to predict decomposed sub-series, and then the obtained sub-series predictions of each individual prediction model are aggregated to generate the preliminary stock price predictions. In the second stage, an ELM-based nonlinear ensemble strategy is employed to combine preliminary stock price predictions. To verify the effectiveness of the proposed two-stage model, it is compared with a total of fourteen models in terms of accuracy evaluation, improvement percentage comparison, and statistical test. The empirical results demonstrate that the proposed two-stage model can obtain better performance than other competitor models. [ABSTRACT FROM AUTHOR]

In recent years, the escalating global concerns surrounding climate change have placed a growing emphasis on achieving dual objectives: reducing carbon emissions and achieving carbon neutrality. Businesses and organizations are under mounting pressure to align their operations with these crucial environmental goals. This paper introduces the concept of an enterprise carbon performance evaluation index system (ECPIS) to strike a balance between economic development and environmental protection and enhance overall enterprise management and development strategies. The ECPIS framework is constructed using machine learning and advanced data mining techniques, particularly support vector machines (SVM). Its core purpose is to provide enterprises with a systematic tool to gauge, analyze, and enhance their carbon performance, addressing dual carbon objectives. ECPIS development hinges on data mining techniques, extracting insights from diverse data sources to construct a comprehensive system that accommodates these dual carbon goals' intricacies. Its methodology includes data collection, preprocessing, feature selection, and data mining algorithms to unveil vital patterns and relationships within data. It conforms to international standards, establishing a tailored carbon performance index system aligned with China's national conditions. It validates carbon-related enterprise data and employs data mining's association rules to uncover pertinent carbon performance information. The results obtained from ECPIS are auspicious, boasting an experiential accuracy rate of 97.5%. This level of accuracy surpasses that achieved by other algorithms like K-Nearest Neighbors (KNN), Random Forest (RF), Decision Trees (DT), Naïve Bayes (NB), and Logistic Regression (LR). ECPIS stands out by considering various factors, including carbon emissions reduction, energy consumption, supply chain efficiency, and financial performance indicators. This multifaceted approach enables enterprises to gain a comprehensive understanding of their carbon performance and identify areas for improvement. [ABSTRACT FROM AUTHOR]

This paper develops an integrated framework to forecast the volatility of crude oil prices by considering the impacts of extreme events (structural breaks). The impacts of extreme events are vital to improving prediction accuracy. Aiming to demonstrate the crude oil price fluctuation and the impacts of external events, this paper employs the complementary ensemble empirical mode decomposition (CEEMD). It decomposes the crude oil price into some constituents at various frequencies to extract a market fluctuation, a shock from extreme events and a long-term trend. The shock from extreme events is found to be the most crucial element in deciding the crude oil prices. Then we combine the iterative cumulative sum of squares (ICSS) test with the Chow test to get the structural breaks and analyze the extreme event impacts. Finally, this paper combines the structural breaks, the autoregressive integrated moving average (ARIMA) model, and the support vector machine (SVM) to make a forecast of the crude oil prices. The empirical process proves that the CEEMD-ARIMA-SVM model with structural breaks performs the best when compared with the other ARIMA-type models and SVM-type models. The framework offers an insightful view to help decision-makers and can be used in many areas. [ABSTRACT FROM AUTHOR]

Multi-view learning improves the performance of existing learning tasks by using complementary information between multiple feature sets. In the latest research, multi-view learning model using privileged information is proposed; specific models are PSVM-2V and MCPK. In these models, views complement each other by acting as privileged information policies; however, a single view contains privileged information that can guide the classifier, and the existing framework does not consider it. In order to use this information to correct multi-view support vector machine classifier, we propose a framework for generating a series of small-scale views based on information hidden in a single view, which extends the original multi-view parallel structure to a hierarchical structure with sub-view mechanism. In this paper, two sub-view learning structures SL-PSVM-2V and SL-MCPK are constructed. The two new models fully exploit the data features in the view. Similarly, they follow the principles of consistency and complementarity. We use the standard quadratic programming solver to solve the new model. In 55 groups of classification experiments, the proposed model improves the classification accuracy by about 1.91% on the original basis. SL-MCPK ranks 1.3846 on average in the accuracy experiment, indicating that they have good classification ability. In addition, the computational time statistics and noise data set experiments are carried out to prove the effectiveness of the proposed method from multiple perspectives. [ABSTRACT FROM AUTHOR]

In the prediction of turning points (TPs) of time series, the improved model of integrating piecewise linear representation and weighted support vector machine (IPLR-WSVM) has achieved good performance. However, due to the single data source and the limitation of algorithm, IPLR-WSVM has encountered challenges in profitability. In this paper, a model fusion method based on multi-source heterogeneous data and different learning algorithms is proposed for the prediction of TPs (MF-MSHD). Multi-source heterogeneous data include weighted unstructured and structured information with different granularities. RF, WSVM, BPNN, GBDT, and LSTM are selected to be the learning algorithms. The differences among meta-models are constructed by different inputs and algorithms as much as possible, and a model fusion rule is designed to determine the final TPs. Moreover, the TPs are generated based on the characteristics of individual stock. For sentiment analysis, a more accurate sentiment dictionary of stock market comments is established. Specifically, the fine-grained data is introduced to jointly determine the accurate trading moment. The prediction level of the proposal improves the accuracy and profitability, and also outperforms the composite indexes. Experimental results show that the profit rate of randomly selected stocks in MF-MSHD reaches 0.5172, while the highest value is 0.2841 in single meta-model and 0.0992 in buy and hold strategy, respectively. The other indicators including the accuracy are also modified. Compared with the increases of 0.1648, 0.4051, and 0.3397 in Shanghai Composite Index, Shenzhen Composite Index, and CSI 300 Index, MF-MSHD shows higher profitability in stock trading signal prediction. [ABSTRACT FROM AUTHOR]

Support vector machine (SVM) is a popular and important machine learning technique based on the central idea of maximizing the minimummargin , i.e., the smallest distance from the training points to the classification boundary. However, recent theoretical studies revealed that maximizing the minimum is not crucial to the generalization performance, while the margindistribution is essential. Furthermore, kernel-free quadratic support vector machine becomes a new trend in the SVM area since it can avoid some difficulties such as finding appropriate kernel functions and tuning their hyper-parameters. This paper proposes two novel non-kernel quadratic surface optimal margin distribution machine (QSODM L and QSODM) models. Their basic idea is to optimize the margin distribution of training points by maximizing the margin mean, minimizing the margin variance and classifying data points by directly constructing a quadratic surface in the original space. These proposed models are convex so that they can be solved by some well-known solvers. Moreover, in order to improve the efficiencies of proposed models on large-scale data sets, a stochastic gradient descent with variance reduction (SVRG) algorithm is proposed to solve the proposed models. The numerical results on some artificial and benchmark data sets verify that QSODM L and QSODM are superior to some other well-known SVM models in classification accuracy and computational efficiency. [ABSTRACT FROM AUTHOR]

Support vector machine (SVM) is a popular supervised machine learning technique extensively applied to various real-life applications. A weakness of SVM, though, is that its efficiency rapidly decreases as the size of training dataset increases. This issue is more challenging for the devices and applications of IoT where the computing resource is quite constrained. As a result, numerous schemes have been proposed to decrease the training time of SVM, especially for the large-scale dataset. In this paper, a survey on the state-of-the-art methods for reducing the number of training data points and increasing the speed of SVM is presented. The existing methods are categorized into five types based on the approach employed for data reduction process, namely clustering-based, geometrical, distance-based, random sampling, and evolutionary methods. The key characteristics of the schemes are reviewed and compared against six practical benchmarks. This survey will be useful to select an appropriate method for the given target problem and develop a new approach for improving the performance of SVM dealing with the large-scale dataset. [ABSTRACT FROM AUTHOR]

Label-specific features learning is a multi-label learning framework that utilizes label feature extraction to solve a single example where multiple class labels exist simultaneously. As an essential multi-label learning method, label correlation learning has been widely used in multi-label classification learning. However, in the existing label-specific features learning, the label correlation measurement often assumes that the label correlations are a global structure or that the label correlations only have a local smoothness. In actual application scenarios, the two situations may occur together. This paper proposes a multi-label classification method by joint Label-Specific features and Global and Local label correlation learning, named LSGL. Firstly, we obtain the weight of the label-specific features of each class label utilizing the l 1 -norm and then learn high-order global label correlation and label local smoothness. By adding manifold regularization terms, we fully utilize the structural relationship between features and labels and mine global and local label association information. These processes are carried out in a unified optimization model, and each part learns and promotes each other. Finally, in the low-dimensional label-specific features representation learning is to carry out multi-label classification learning through the support vector machine and the extreme learning machine, respectively. A comparative study with state-of-the-art approaches and statistical hypothesis testing manifests the validity of the LSGL method and the features learned from label-specific features learning. [ABSTRACT FROM AUTHOR]

Fuzzy parameterized fuzzy soft matrices (fpfs-matrices) which can model problems involving fuzzy objects and parameters are one of the mathematical tools used to deal with decision-making problems. To utilize soft decision-making methods via fpfs-matrices in machine learning is likely to draw much scholarly attention. In this paper, we propose Comparison Matrix-Based Fuzzy Parameterized Fuzzy Soft Classifier (FPFS-CMC) in order to transfer modeling success of fpfs-matrices to machine learning. We then compare FPFS-CMC with Fuzzy Soft Set Classifier (FSSC), FussCyier, Fuzzy Soft Set Classification Using Hamming Distance (HDFSSC), and Fuzzy k-Nearest Neighbor (Fuzzy kNN) in consideration of accuracy, precision, recall, macro-F-score, and micro-F-score performance metrics, and 15 datasets in UCI Machine Learning Repository. Besides, we compare the proposed classifier with the state-of-the-art Support Vector Machine (SVM), Decision Tree (DT), and Adaptive Boosting (AdaBoost) in terms of five performance metrics herein. Afterward, the results from the experiments are analyzed by employing the Friedman and Nemenyi tests to assess the statistical significance of the differences in performances. Both experimental and statistical results show that FPFS-CMC outperforms the others. Finally, we provide the conclusive remarks and some suggestions for further research. [ABSTRACT FROM AUTHOR]

Fuzzy support vector machine and their variants are prominent classification techniques that reduce the adverse effects of noises and outliers as compare to classical support vector machine. However, in these methods, only the distance between the training pattern and the class center is considered, and hence, the edge support vectors cannot be distinguished from noises and outliers properly. These limitations are overcome by intuitionistic fuzzy-based support vector machine which allocate two parameters (membership and non-membership degrees) to each pattern of a dataset and hence define score number based on the importance of a pattern. In this paper, an intuitionistic fuzzy proximal support vector machine for multi-category classification problems is proposed. The method significantly reduces the impacts of noises and outliers present in the dataset by assigning the intuitionistic fuzzy score function to each training point based on its location and surroundings. Moreover, the method is computationally efficient as the robust classifiers are obtained by solving the system of linear equations instead of large size quadratic programming problems. In the proposed method, using polynomial and Gaussian kernels, the hyperplanes are also developed in the feature space. The geometrical advantages of the suggested method over the existing techniques are ascertained using the simulated two-dimensional artificial dataset having three target classes. Further, extensive experimental studies on ten UCI benchmark datasets have been performed which demonstrate that the proposed algorithm predicts more precisely about future data as comparison to some well-established algorithms. Figures are also illustrated by varying different parameters involved in the model which confirms the performance of the method over the existing algorithms. Moreover, the analysis of the predictive behavior of the proposed approach is done using the Friedman test, a nonparametric alternative to the analysis of variance test at 5 % significance level. Further, the proposed method has also been applied to image classification and gesture-phase segmentation problems which confirms the efficiency and handling capabilities of the proposed algorithm in practical applications. [ABSTRACT FROM AUTHOR]

Almost all of the automatic accident detection (AAD) system suffers from the tradeoff between computational overhead and detection accuracy. Recent advances in detection and classification methodologies have shown phenomenal improvements in accuracy but these systems require a huge number of computational resources making them unviable for deployment requiring real-time feedback. This paper proposes a methodology to develop a reliable and computationally inexpensive real-time automatic accident detection system that can be deployed with minimum hardware requirements. Specifically, we split our AAD system into three major stages (Detection, Tracking and Classification) and propose algorithms for each stage with reduced computational need. For the detection stage, we propose Mini-YOLO, a deep learning model architecture trained using knowledge distillation that has comparable accuracy with its counterpart YOLO(You-Only-Look-Once) with reduced model size and computational overhead. Mini-YOLO achieves an average precision (AP) score of 34.2 on the MS-COCO dataset while outperforming all other detection algorithms in runtime complexity, achieving a staggering 28 frames per second on a low-end machine. For the tracking stage, we adopt SORT (Simple Online Real-time Tracking) and for classification stage, we compare multiple machine learning algorithms and show that a support vector machine with radial basis kernel performs the best with an area under the curve (AUC) score of 0.98, model size of 448 KB (kilobytes) and 12.73 ms (milliseconds) latency. [ABSTRACT FROM AUTHOR]

The existence of both uncertainty and imprecision has detrimental impact on efficiency of decision-making applications and some machine learning methods, in particular support vector machine in which noisy samples diminish the performance of SVM training. Therefore, it is important to introduce a special method in order to improve this problem. Fuzzy aspects can handle mentioned problem which has been considered in some classification methods. This paper presents a novel weighted support vector machine to improve the noisy sensitivity problem of standard support vector machine for multiclass data classification. The basic idea is considered to add a weighted coefficient to the penalty term Lagrangian formula for optimization problem, which is called entropy degree, using lower and upper approximation for membership function in fuzzy rough set theory. As a result, noisy samples have low degree and important samples have high degree. To evaluate the power of the proposed method WSVM-FRS (Weighted SVM-Fuzzy Rough Set), several experiments have been conducted based on tenfold cross-validation over real-world data sets from UCI repository and MNIST data set. Experimental results show that the proposed method is superior than the other state-of-the-art competing methods regarding accuracy, precision and recall metrics. [ABSTRACT FROM AUTHOR]

A noise-aware version of support vector machines is utilized for feature selection in this paper. Combining this method and sequential backward search (SBS), a new algorithm for removing irrelevant features is proposed. Although feature selection methods in the literature which utilize support vector machines have provided acceptable results, noisy samples and outliers may affect the performance of SVM and feature selections method, consequently. Recently, we have proposed relaxed constrains SVM (RSVM) which handles noisy data and outliers. Each training sample in RSVM is associated with a degree of importance utilizing the fuzzy c-means clustering method. Therefore, a less importance degree is assigned to noisy data and outliers. Moreover, RSVM has more relaxed constraints that can reduce the effect of noisy samples. Feature selection increases the accuracy of different machine learning applications by eliminating noisy and irrelevant features. In the proposed RSVM-SBS feature selection algorithm, noisy data have small effect on eliminating irrelevant features. Experimental results using real-world data verify that RSVM-SBS has better results in comparison with other feature selection approaches utilizing support vector machines. [ABSTRACT FROM AUTHOR]

Security is one of the main requirements of the current computer systems, and recently it gains much importance as the number and severity of malicious attacks increase dramatically. Anomaly detection is one of the main branches of the intrusion detection systems which enables to recognize the newer variants of the security attacks. This paper focuses on the anomaly detection schemes (ADS), which have applied support vector machine (SVM) for detecting intrusions and security attacks. For this purpose, it first presents the required concepts about the SVM classifier and intrusion detection systems. It then classifies the ADS approaches and discusses the various machine learning and artificial intelligence techniques that have been applied in combination with the SVM classifier to detect anomalies. Besides, it specifies the primary capabilities, possible limitations, or advantages of the ADS approaches. Furthermore, a comparison of the studied ADS schemes is provided to illuminate their various technical details. [ABSTRACT FROM AUTHOR]

In numerous real-world problems, we are faced with difficulties in learning from imbalanced data. The classification performance of a "standard" classifier (learning algorithm) is evidently hindered by the imbalanced distribution of data. The over-sampling and under-sampling methods have been researched extensively with the aim to increase the predication accuracy over the minority class. However, traditional under-sampling methods tend to ignore important characteristics pertinent to the majority class. In this paper, a novel under-sampling method based on information granules is proposed. The method exploits the concepts and algorithms of granular computing. First, information granules are built around the selected patterns coming from the majority class to capture the essence of the data belonging to this class. In the sequel, the resultant information granules are evaluated in terms of their quality and those with the highest specificity values are selected. Next, the selected numeric data are augmented by some weights implied by the size of information granules. Finally, a support vector machine and a K-nearest-neighbor classifier, both being regarded here as representative classifiers, are built based on the weighted data. Experimental studies are carried out using synthetic data as well as a suite of imbalanced data sets coming from the public machine learning repositories. The experimental results quantify the performance of support vector machine and K-nearest-neighbor with under-sampling method based on information granules. The results demonstrate the superiority of the performance obtained for these classifiers endowed with conventional under-sampling method. In general, the improvement of performance expressed in terms of G-means is over 10% when applying information granule under-sampling compared with random under-sampling. [ABSTRACT FROM AUTHOR]

Due to complex nature of nearly all infrastructures (and more specifically concrete dams), the uncertainty quantification is an inseparable part of risk assessment. Uncertainties might be propagated in different aspects depending on their relative importance such as epistemic and aleatory, or spatial and temporal. The objective of this paper is to focus on the material and modeling uncertainties, and to couple them with soft computing techniques aiming to reduce the computational burden of the conventional Monte Carlo-based finite element simulations. Several scenarios are considered in which the concrete and foundation material properties, the water level, and the dam geometry are assumed as random variables. Five soft computing techniques (i.e., random forest, boosted regression trees, multi-adaptive regression splines, artificial neural networks, and support vector machines) are employed to predict various quantities of interest based on different training sizes. It is argued that the artificial neural network is the most accurate algorithm in majority of cases, with enough accuracy as to be useful in reliability analysis as a complement to numerical models. The results with 200 samples in the training set are enough for reaching useful accuracy in most cases. For the simple prediction tasks, the results were predicted with less than 1% error. It is observed that increasing the number of input parameters increases the prediction error. The partial dependence plots provided most sensitive variables in dam design, which were consistent with the physics of the problem. Finally, several practical recommendations are provided for future applications. [ABSTRACT FROM AUTHOR]

Machine learning methods are increasingly used in proteomics research, especially in analyzing and predicting protein structures, functions, subcellular localizations and interactions. However, much research in recent years has focused on protein misfolding problem and the impact of unfolded and defective proteins on cell dysfunction, due to its considerable importance for molecular medicine. These abnormal proteins degradation and deposition often result in the formation of certain plaque cores among them the so-called amyloid fibrils which are responsible for an increasing number of highly debilitating disorders in humans. Yet, a significant challenge remains, especially in understanding the underlying causes and major risk factors of these harmful deposits in vital organs and tissues. This paper explores the potential of string kernel-based support vector machines in the prediction of amyloidogenic regions in proteins by incorporating the most informative features of the protein sequence such as predicted secondary structure and solvent accessibility, with a special focus on α -helical conformations which seem to be primarily concerned with amyloidogenesis. The performances compared with the most popular methods on Pep424 and Reg33 benchmark datasets indicate the robustness of the predictive model. Furthermore, the results showed accurate prediction of regions promoting fibrillogenesis for experimentally determined amyloid proteins and revealed that the five amino acids Leucine, Glycine, Alanine, Valine and Serine are predominantly present in amyloid-prone regions and confirm that the core regions of an amyloid aggregate are not necessarily fully buried. [ABSTRACT FROM AUTHOR]

Fuzzy neural network (FNN) and support vector machine (SVM) are two prominent and powerful learning models broadly used for classification and regression. FNN has a significant local representation and human reasoning advantage. However, the drawback of such a network is that the focal point of the learning algorithms is minimizing empirical risk. In contrary to FNN, SVM emphasizes simultaneously on minimizing empirical and expected risks, which theoretically leads to an excellent generalization performance power. In this paper, we show that a Takagi–Sugeno–Kang (TSK)-type-based fuzzy neural network is, in fact, equivalent to an SVM with an adaptive kernel based on fuzzy rules generated in this FNN. Consequently, it is possible to learn the last layer of the FNN using the concepts of SVM and thus taking the advantage of SVM in generalization. In fact, on the one side, the proposed method is an SVM with an adaptive kernel based on fuzzy rules and, on the other side, it is a TSK–FNN with SVM-based learning. As a matter of fact, the defined kernel in FNN is an adaptive kernel based on data characteristics, which is derived from the fuzzy rules generated by the FNN itself. The results obtained from the proposed method for classification and regression outperform the results of SVM with the conventional kernels and fuzzy neural network trained in the traditional way. [ABSTRACT FROM AUTHOR]

Decisions made by deans and university managers greatly impact the entire academic community as well as society as a whole. In this paper, we present survey results on which academic decisions they concern and the variables involved in them. Using machine learning algorithms, we predicted graduation rates in a real case study to support decision making. Real data from five undergraduate engineering programs at District University Francisco Jose de Caldas in Colombia illustrate our results. The comparison between support vector machine and artificial neural network is held using the confusion matrix and the receiver operating characteristic curve. The algorithm methods and architecture are presented. [ABSTRACT FROM AUTHOR]

Local multiple kernel learning is a promising strategy because it could learn a sample-specific composite kernel according to the characteristic of samples. However, these methods are insufficient to describe the sample diversity and correlation, which leads to the classifier less reliable. In this paper, we propose a soft-clustering-based local multiple kernel learning algorithm to tackle the issues above. In the proposed algorithm, there is a fuzzy clustering preprocessing for the training data and then the kernel weights are calculated on the groups. We use an alternative optimization method to learn the kernel weights and support vector coefficients. The final combination weights of kernels are determined by the kernel weights of clusters and the probability of samples falling into the clusters. Therefore, our method is actually a sample-based LMKL method with a soft constraint on the kernel weights. This constraint is actually the representation of the correlation of samples. The experiments on synthetic dataset indicate the kernel weights solved by our algorithm are better suitable for the characteristics of the dataset. Then a series of experiments verify the improvement on classification accuracies. [ABSTRACT FROM AUTHOR]

In this paper, an enhanced quantum-based neural network learning algorithm (EQNN-S) which constructs a neural network architecture using the quantum computing concept is proposed for signature verification. The quantum computing concept is used to decide the connection weights and threshold of neurons. A boundary threshold parameter is introduced to optimally determine the neuron threshold. This parameter uses min, max function to decide threshold, which assists efficient learning. A manually prepared signature dataset is used to test the performance of the proposed algorithm. To uniquely identify the signature, several novel features are selected such as the number of loops present in the signature, the boundary calculation, the number of vertical and horizontal dense patches, and the angle measurement. A total of 45 features are extracted from each signature. The performance of the proposed algorithm is evaluated by rigorous training and testing with these signatures using partitions of 60–40 and 70–30%, and a tenfold cross-validation. To compare the results derived from the proposed quantum neural network, the same dataset is tested on support vector machine, multilayer perceptron, back propagation neural network, and Naive Bayes. The performance of the proposed algorithm is found better when compared with the above methods, and the results verify the effectiveness of the proposed algorithm. [ABSTRACT FROM AUTHOR]

Semi-supervised support vector machine is a popular method in the research area of machine learning. Considering a large amount of unlabeled data points in real-life world, the semi-supervised support machine has the ability of good generalization for dealing with nonlinear classification problems. In this paper, a proximal quadratic surface support vector machine model is proposed for semi-supervised binary classification. The main advantage of our new model is that the proximal quadratic surfaces are constructed directly for nonlinear classification instead of using the kernel function, which avoids the tasks of choosing kernels and tuning their parameters. We reformulate this proposed model as an unconstrained mixed-integer quadratic programming problem. Semi-definite relaxation is then adopted, and a primal alternating direction method is further proposed for fast computation. We test the proposed method on some artificial and public benchmark data sets. Preliminary results indicate that our method outperforms some well-known methods for semi-supervised classification in terms of the efficiency and classifying accuracy. [ABSTRACT FROM AUTHOR]

Most research on multi-document summarization explores methods that generate summaries based on queries regardless of the users' preferences. We note that, different users can generate somewhat different summaries on the basis of the same source data and query. This paper presents our study on how to exploit the information regards how users summarized their texts. Models of different users can be used either separately, or in an ensemble-like fashion. Machine learning methods are explored in the construction of the individual models. However, we explore yet another hypothesis. We believe that the sentences selected into the summary should be coherent and supplement each other in their meaning. One method to model this relationship between sentences is by detecting actor-object relationship (AOR). The sentences that satisfy this relationship have their importance value enhanced. This paper combines ensemble summarizing system and AOR to generate summaries. We have evaluated this method on DUC 2006 and DUC 2007 using ROUGE measure. Experimental results show the supervised method that exploits the ensemble summarizing system combined with AOR outperforms previous models when considering performance in query-based multi-document summarization tasks. [ABSTRACT FROM AUTHOR]

In this paper, a novel data selection algorithm for identifying significant training data is presented for power quality (PQ) events classification. The key concept of this paper is to reduce the execution time, computational complexity and enhance the accuracy of the existing PQ classification system by reducing the number of support vectors. The proposed scheme identifies important training data and rejects redundant and irrelevant ones using a novel fuzzy C-means clustering-based data selection algorithm thereby reducing time of classification and enhancing accuracy. Significant features from raw PQ data are extracted using discrete wavelet transform and important training data are recognized using these features with the proposed data selection algorithm. Two machine learning algorithms namely the probabilistic neural network and support vector machine are employed and the best PQ classifier is investigated. Furthermore, the proposed data selection algorithm is integrated with the existing PQ classification system that has selected optimal input features and parameters of the classifier using Simulated Annealing and is shown to perform exceptionally well when compared to conventional classifiers that use full training data set. The suitability of the algorithm for noisy as well as real time data is examined and the empirical results show that the proposed scheme performs well compared to several existing PQ classification works. [ABSTRACT FROM AUTHOR]

Learning from categorical data is a critical yet challenging task. Current research focuses on either leveraging the complex interaction between and within categorical values to generate a numerical representation, or designing a model that can tackle this types of data directly. However, both of these paradigms overlook the relation between the data characteristics and learning model hypothesis. In this paper, we propose a model-aware categorical data embedding framework that jointly reveals the intrinsic categorical data characteristics and optimizes the fitness of the representation for the follow-up learning model. An ELM-aware and a SVM-aware representation methods have been instantiated under this framework. Extensive experiments of classification with the embedded representation on 17 data sets demonstrate that the proposed framework can significantly improve the categorical data representation performance compared with state-of-the-art competitors. [ABSTRACT FROM AUTHOR]

Large-scale problems have been a very active topic in machine learning area. In the time of big data, it is a challenge and meaningful work to solve such problems. Standard SVM can make linear classification on large-scale problems effectively, with acceptable training time and excellent prediction accuracy. However, nonparallel SVM (NPSVM) and ramp loss nonparallel SVM (RNPSVM) are proposed with better performance than SVM on benchmark datasets. It is motivated to introduce NPSVMs into the area of large-scale issues. In this paper, we propose large-scale linear NPSVMs, solved by the alternating direction method of multipliers (ADMM), to handle large-scale classification problems. ADMM breaks large problems into smaller pieces, avoiding solving intractable problems and leading to higher training speed. The primal problems of NPSVM are convex and differentiable, and they can be managed directly by ADMM. But the objective functions of RNPSVM, composed of convex ones and concave ones, should first be processed by CCCP algorithm and transformed as a series of convex programs. Then, we apply ADMM to solve these programs in every iteration. Experiments of NPSVMs on large-scale problems verify that the algorithms can classify large-scale tasks effectively. [ABSTRACT FROM AUTHOR]

This paper reports an exhaustive analysis performed over two specific Genetics-based Machine Learning systems: BioHEL and GAssist. These two systems share many mechanisms and operators, but at the same time, they apply two different learning paradigms (the Iterative Rule Learning approach and the Pittsburgh approach, respectively). The aim of this paper is to: (a) propose standard configurations for handling small and large datasets, (b) compare the two systems in terms of learning capabilities, complexity of the obtained solutions and learning time, (c) determine the areas of the problem space where each one of these two systems performs better, and (d) compare them with other well-known machine learning algorithms. The results show that it is possible to find standard configurations for both systems. With these configurations the systems perform up to the standards of other state-of-the-art machine learning algorithms such as Support Vector Machines. Moreover, we identify the problem domains where each one of these systems have advantages and disadvantages and propose ways to improve the systems based on this analysis. [ABSTRACT FROM AUTHOR]

In this paper, we propose a novel multi-parametric kernel Support Vector Regression algorithm (SVMr) optimized with an evolutionary technique, specially well suited for forecasting problems. The multi-parametric SVMr model and the evolutionary algorithm proposed are both described in detail in the paper. In addition, several new bounds for the multi-parametric kernel considered are obtained, in such a way that the SVMr hyper-parameters' search space is reduced. We present experimental evidences of the good performance of the evolutionary algorithm for optimizing the multi-parametric kernel, when compared to a standard SVMr with a Grid Search approach. Specifically, results in different real regression problems from public repositories are obtained, and also a real application focused on the short-term temperature prediction at Barcelona's airport. The results obtained have shown the good performance of the multi-parametric kernel approach both in accuracy and computation time. [ABSTRACT FROM AUTHOR]

Stock price volatility prediction is regarded as one of the most attractive and meaningful research issues in financial market. Some existing researches have pointed out that both the prediction accuracy and the prediction speed are the most important factors in the process of stock prediction. In this paper, we focus on the problem of how to design a methodology which can improve prediction accuracy as well as speed up prediction process, and propose a new prediction model which employs mutual information- based sentimental analysis methodology with extreme learning machine to enhance the prediction performance. The two major contributions of our work are (1) as the words in the news documents are not absolutely negative or positive, and the lengths of the financial news documents are various; here, we propose a new sentimental analysis methodology based on mutual information to improve the efficiency of feature selection, which is different from the traditional sentimental analysis algorithm, and a new weighting scheme is also used in the feature weighting process; (2) since ELM is a fast learning model and has been successfully applied in many research fields, we propose a prediction model which combined mutual information-based sentimental analysis with kernel-based ELM named as MISA-K-ELM. This model has the benefits of both statistical sentimental analysis and ELM, which can well balance the requirements of both prediction accuracy and prediction speed. We take experiments on HKEx 2001 stock market datasets to validate the performance of the proposed MISA-K-ELM. The market historical price and the market news are implemented in our MISA-K-ELM. To test the efficiency of MISA, we first compare the prediction accuracy of ELM model using MISA with ELM model using traditional sentimental analysis. Then, we compare our proposed MISA-K-ELM with existing state-of-the-art learning algorithms, such as Back-Propagation Neural Network (BP-NN), and Support Vector Machine (SVM). Our experimental results show that (1) MISA model can help get higher prediction accuracy than traditional sentimental analysis models; (2) MISA-K-ELM and MISA-SVM have a higher prediction accuracy than MISA-BP-NN and MISA-B-ELM; (3) both MISA-K-ELM and MISA-B-ELM can achieve faster prediction speed than MISA-SVM and MISA-BP-NN in most cases; (4) in most cases, MISA-K-ELM has higher prediction accuracy than the other three methodologies. [ABSTRACT FROM AUTHOR]

Functional data learning is an extension of traditional data learning, that is, learning the data chosen from the Euclidean space $${\mathbb{R}^{n}}$$ to a metric space. This paper focuses on the functional data learning with generalized single-hidden layer feedforward neural networks (GSLFNs) acting on some metric spaces. In addition, three learning algorithms, named Hilbert parallel overrelaxation backpropagation (H-PORBP) algorithm, ν-generalized support vector regression (ν-GSVR) and generalized extreme learning machine (G-ELM) are proposed to train the GSLFNs acting on some metric spaces. The experimental results on some metric spaces indicate that GELM with additive/RBF hidden-nodes has a faster learning speed, a better accuracy, and a better stability than HPORBP algorithm and ν-GSVR for training the functional data. The idea of GELM can be used to extend those improved extreme learning machines (ELMs) that act on the Euclidean space $${\mathbb{R}^{n}, }$$ such as online sequential ELM, incremental ELM, pruning ELM and so on, to some metric spaces. [ABSTRACT FROM AUTHOR]

Extreme learning machine (ELM), as an emergent technique for training feed-forward neural networks, has shown good performances on various learning domains. This paper investigates the impact of random weights during the training of ELM. It focuses on the randomness of weights between input and hidden layers, and the dimension change from input layer to hidden layer. The direct motivation is to verify as to whether during the training of ELM, the randomly assigned weights exert some positive effects. Experimentally we show that for many classification and regression problems, the dimension increase caused by random weights in ELM has a performance better than the dimension increase caused by some kernel mappings. We assume that via the random transformation, output-samples are more concentrate than input-samples which will make the learning more efficient. [ABSTRACT FROM AUTHOR]

A human face does not play its role in the identification of an individual but also communicates useful information about a person's emotional state at a particular time. No wonder automatic face expression recognition has become an area of great interest within the computer science, psychology, medicine, and human-computer interaction research communities. Various feature extraction techniques based on statistical to geometrical data have been used for recognition of expressions from static images as well as real-time videos. In this paper, we present a method for automatic recognition of facial expressions from face images by providing discrete wavelet transform features to a bank of seven parallel support vector machines (SVMs). Each SVM is trained to recognize a particular facial expression, so that it is most sensitive to that expression. Multi-classification is achieved by combining multiple SVMs performing binary classification using one-against-all approach. The outputs of all SVMs are combined using a maximum function. The classification efficiency is tested on static images from the publicly available Japanese Female Facial Expression database. The experiments using the proposed method demonstrate promising results. [ABSTRACT FROM AUTHOR]

Abstract  Aiming at the common support vector machine’s biased disadvantage and computational complexity, an unbiased least squares support vector machine (LSSVM) model is proposed in this paper. The model eliminates the bias item of LSSVM by improving the form of structure risk, then the unbiased least squares support vector classifier and the unbiased least squares support vector regression are deduced. Based on this model, we design a new learning algorithm using Cholesky factorization according to the characteristic of kernel function matrix, in this way the calculation of Lagrangian multipliers is greatly simplified. Several experiments on diffenert datasets are carried out, including the common datasets classification, synthetic aperture radar image automatic target recognition and chaotic time series prediction. The experimental results of correct recognition rate and the fitting precision testify that the unbiased LSSVM model has good universal ability and fitting accuracy, better generalization capability and stability, and have a great improvement in learning speed. [ABSTRACT FROM AUTHOR]