Search

Clustering is one of the popular methods for improving energy efficiency in wireless sensor networks. In most of the existing fuzzy approaches, the CHs are selected first, and then clusters are generated, but this may lead to uneven distribution of the sensor nodes in the clusters. In this article, the clusters are generated using the famous Fuzzy C-means (FCM) algorithm and the Cluster Head (CH) from each cluster is selected using the Sugeno fuzzy system. FCM generates load-balanced clusters and the proposed approach named SF-MPSO selects the suitable CH from each cluster. The local information of the sensor node such as residual energy, its distance from cluster centroid and the distance from the BS is provided to SF-MPSO. In the existing algorithms, the fuzzy rules are manually designed, whereas, in this article, the modified Particle Swarm Optimization (PSO) algorithm is applied to generate optimum Sugeno fuzzy rules. A novel fitness function is designed to identify the effectiveness of the generated solution. The simulations are performed under three scenarios where SF-MPSO outperforms existing EAUCF, DUCF, FGWO and ARSH-FATI-CHS when evaluated under the parameters such as energy consumption and network lifetime.

The weather forecasting process is used to predict the future atmospheric condition of a specified location. The evolution of the big data era gives the chances to significantly increase the prediction accuracy of weather conditions. In this paper, a deep learning-based stacked sparse autoencoder (DSSAE) has been proposed for predicting the weather condition of a particular area. This model requires a pre-processing approach to obtain essential data from big weather data and increase the prediction model’s speed. For this, the principal component analysis (PCA) is utilized to reduce dimensionality and extraction of features with more significant variance. Also, it integrates the feature selection algorithm based on Binary Butterfly Optimization Algorithm (BBOA) along with a deep stack autoencoder to improve prediction accuracy. The proposed model is validated using the weather data taken from the division of weather underground for short term and long term weather prediction. The simulation consequences illustrate that the proposed model overtakes the existing models in terms of computation time, accuracy and error rate.

The increasing population in Asia brings up the need for integration of healthcare for efficient and timely manageable treatment for different diseases. Healthcare domain is one of the most important and challenging fields in terms of data collection and analysis. This domain always provide lots of opportunities to explore the hidden knowledge in accessing health records. With the growth of unstructured data in large volume that leads towards the solution by the NoSQL data management tool to manage the huge amount of data. This framework proposes a MapReduce Approach (MRA) for data management in healthcare industry with join based expectation maximization algorithm for NoSQL data management solution, which scales the data with accurate modality. This approach also simplifies the way to integrate healthcare data from different models in the distributed environment from different health hubs. Experimental results show that the proposed approach works in a scalable manner to integrate and match the unstructured data of different health data sources. Examples are illustrated with suitable methodology and further research scope is pinpointed.

NoSQL database technology has been doing rounds since the early 1990s, but it was the exponential growth of internet and the rise of web applications that lead to a dynamic surge in the popularity of NoSQL databases. The BigTable research by Google (2006) and the Dynamo research by Amazon (2007) paved the way for databases which could develop with agility and operate at any scale. Cassandra and MongoDB have emerged as the two most widely used NoSQL database and hence either of the two is preferred depending on the data problem user is attempting to solve. This paper describes the underlying principles as well as the differences between both the databases. We focus on showing the anomaly in performance of Cassandra as the data volume increases and at the same time we compare its performance with that of MongoDB. We establish how important factor is data volume in choosing either of the databases for an application. Extensive experiments have been carried out to scale the performance in terms of anomaly similarities, and the future scope is pinpointed.

Visualization of data which is enormously generated at present has become one of the major steps in machine learning before applying the classifier on related domain problems. This paper inculcates the visualization and recognition of human activity from n-dimensional space to low-dimensional space by implementing an algorithm and sometimes best termed as a tool called as t-SNE (t-distributed stochastic neighborhood embedding), which is a non-linear algorithm and has been inspired by SNE (stochastic neighbor embedding). This algorithm is applied to build the relationship between various actions done by human, where class label information has been introduced to well represent the actions from similar action class. This comes under a machine learning methodology called manifold learning, where the class label information helps to classify the action done by human. This learning technique uses probabilistic embedding of neighbors from n-dimensional space with low-dimension to find a mapping of data points from one distribution to another distribution. Experimental results and their comparison with other visualization tools like PCA shows the effectiveness of t-SNE algorithm with human activity recognition and MNIST dataset. The results produced by t-SNE algorithm in terms of visualization are always better than many other algorithms.

With the exponential growth of the Internet, the need of handling the sort of unstructured, messy and unpredictable data became so urgent. This makes NoSQL databases to be introduced to compensate for the inefficiency of traditional relational databases in dealing with these kind of data. Among those NoSQL databases, Cassandra gained popularity than the rest and today it is at use in over 1500 companies like Reddit, Netflix, Instagram, GitHub, eBay. The EQUALS and IN operator are the two most basic tools used in Cassandra Query Language (CQL) to compare two or more data items. However, the two operators are not capable of dealing with non-traditional data types like images as instead of looking for similarity between two images, they use identity (=) which is not always the correct method. Thus, this paper introduces similarity-aware EQUALS and IN operator which is very much capable of dealing with modern complex data types like images. These operators based on similarity can be used in various fields like Agriculture for finding regions capable of growing a particular crop. In this paper, we demonstrate the application of the two operators in Agriculture.

Fuzzy classification rules are more interpretable and cope better with pervasive uncertainty and vagueness with respect to crisp rules. Because of this fact, fuzzy classification rules are extensively used in classification and decision support systems for disease diagnosis. But, most of the rule mining techniques failed to generate accurate and comprehensive fuzzy rules. This paper presents a hybrid decision support system based on Rough Set Theory (RST) and Bat optimization Algorithm (BA) called RST-BatMiner. It consists of two stages. In the first stage, redundant features have been removed from the data set through RST-based QUICK-REDUCT approach. In the second stage, for each class BA is invoked to generate fuzzy rules by minimizing proposed fitness function. Further, an Ada-Boosting technique is applied to the rules generated by BA to increase the accuracy rate of generated fuzzy rules. Moreover, to generate comprehensive fuzzy rules, a new ≠ (not equal) operator along with = (equal) operator is introduced into BA encoding scheme. The proposed RST-BatMiner builds consolidated fuzzy ruleset by learning the rules associated with each class separately. The proposed RST-BatMiner is experimented on six bench-mark datasets namely Pima Indians Diabetes, Wisconsin Breast Cancer, Cleveland Heart disease, iris, wine and glass, in order to validate its generalization capability. These experimental results show that except for wine dataset the proposed RST-BatMiner yields high accuracy and comprehensible ruleset when compared to other state-of-the-art bio-inspired based fuzzy rule miners and Fuzzy Rule Based Classification Systems (FRBCS) in the literature. In the case of wine dataset, the proposed RST-BatMiner yields second highest accuracy along with comprehensible ruleset.

Link prediction in a given instance of a network topology is a crucial task for extracting and inspecting the evolution of social networks. It predicts missing links in existing community networks and new or terminating links in future systems. It also attracted much attention in many fields. In the past decade, many methodologies have been compiled to predict the suitable links in a given social network. Analyzing link prediction methods is difficult when the network is very complex due to restrictive computing cost. It is still a very challenging task to predict missing links efficiently and accurately in an incomplete complex network. Depending on the certainty, the nodes with an incredible number of normal neighbors will probably be connected. Numerous similarity indices have accomplished extensive exactness and efficiency that greatly optimized this task. To accommodate this instance, in this paper, we propose one such index, namely Clustering Coefficient Index, using triangle counting implemented on the component of Apache Spark’s GraphX methodology. The proposed index uses the property of formation of triangles in the given network topology and clustering coefficients. Experimental results show that the proposed methodology outperforms in linking the suitable communications compared to other existing methods.

Brain computer interface (BCI) provides communication between the computer and the brain. It is the combination of hardware and software which provides non-muscular channel to send the various messages to control the computer. BCI is useful in various medical applications such as patients with neuromuscular injuries, locked-in syndrome (LiS) etc. BCI is not only useful in medical applications, but also useful in lie detection, entertainment, etc. In this paper, spark and rule-KNN based scalable framework has been presented using BCI with the EEG data collected on 20 subjects in which 10 are acted as innocent and 10 are acted as guilty. Using BCI P300, Deceit identification Test (DIT) is performed. To perform DIT, we classify the P300 signals which have a positive peak of 300 ms–1000 ms in one stimulus start. Data processing is performed with band pass filter to cut the frequency ranges and features are extracted using non-parametric weighted feature extraction followed by rule based discriminant classification. For training and testing, the data ratio selected as 80:20 and achieved the accuracy 92.46 %. Proposed framework provides better results in comparison with existing models presented in literature. Hence this model is accurate, scalable and fault tolerant.

In the current digital era, huge data are being generated in a voluminous state from different sources. This lead towards a processing repository called Big Data. Managing and processing such data in parallel clusters is a big challenge. To capture this problem, in this paper, we propose a hybrid algorithm for cluster analysis using the Spark framework for analyzing the Big Data instances. The proposed algorithm is the combination of two machine learning techniques namely, K-Means (KM) and C5.0 Decision Tree (DT). As per the factor of cluster, euclidean distance is used to find the nearest cluster and the related DT is built for each cluster using C5.0 DT algorithm. The inferences of the DT are used to classify each anomaly and the normal instances of the large datasets. Experimental results show that the proposed hybrid algorithm outperforms with other existing algorithms and produces better classification accuracy for anomaly detection.

First, we expose the instance of ER with EM algorithm to integrate different models.We apply ER for HDPD to achieve major performance in terms of response time.Matching instance is compared with BFA, Koosh-TND and Packages-TND algorithms.After the integration, we work on real data and compare with FC, SCC and FD algorithms.Our method improves in terms of response and communication time with existing methods. Distributed computing is linked and equated to the industrial revolution. Its transformational nature is, however, associated with significant instances in the form of internet of thing operations. Entity resolution (ER) is a problem of matching and resolving records that represent the same real world entity. This is a long-standing challenge in distributed databases and information retrieval as a statistic. In a centralized approach, the problem of ER has not been scaled well as large amount of data need to be sent to a central node. In this paper, we present an algorithm which deals with heterogeneous distributed probabilistic data (HDPD) and also reduces processing time in a distributed environment. We propose two different approaches. First, we explore this instance with a matching (identification) problem to integrate different data models with expectation-maximization (EM) algorithm. Second, we apply ER methodology for HDPD to achieve major performance in terms of response time to produce the outcome. We validate HDPD through experiments over a 100-node cluster that records significant performance improvements over naive approaches. This paper is expected to provide insights in to database organizations and new technological development for the growth of distributed environment.

With the development of cloud computing and internet; e-Commerce, e-Business and corporate world revenue are increasing with high rate. These areas not only require scalable and consistent databases but also require inter database transaction support. In this paper, we present, a scalable three-tier architecture along with a distributed middle-ware protocol to support atomic transactions across heterogeneous NoSQL databases. Our methodology does not compromise on any assumption on the accuracy of failure modalities. Hence, it is suitable for a class of heterogeneous distributed systems. To achieve such a target, our architectural model exploits an innovative methodology to achieve distributed atomic transactions. We simulate this architectural setup with different latency tests under different environments to produce reliable impact and correctness. This study proposes a three-tier architecture to facilitate atomic transactions among heterogeneous NoSQL distributed database environment as HBase.The first tier called client tier, deals with user requests. The client tier wraps transaction request and deliver it to a corresponding global transaction manager (GTM), which is assigned by a manager node.The second tier called middle tier, contains a zookeeper, manager nodes, and labour nodes. Zookeeper assigns an active manager node to the client and the active manager node assigns a labour node to process the transaction.The third tier called distributed tier manages multiple heterogeneous NoSQL databases participating with different transactions.Experimental results show that the proposed architecture is scalable and fault tolerated.

Association rule mining is one of prominent techniques to discover the relation between data items of a transactional data. The process of mining has been simplified by considering only the frequent itemsets. Pincer search is one of the frequent itemset mining method which combines top-down and bottom-up search techniques to get the benefits of both. Top-down approach in Pincer search reduces the number of candidates in pass of iterations and saves a lot of computing resources. In this work, we present a Parallel Pincer Search (PPS) which is based on distributed implementation on Spark framework. We have converted the search algorithm according to the Spark framework to make it run in parallel. Spark provides a lot of features for the iterative algorithm such as in-memory execution, efficient data structure, better fault tolerant method, etc. We implemented the PPS on a Spark cluster with multiple datasets and analysed the performance.

Low-power wearable devices for disease diagnosis are used at anytime and anywhere. These are non-invasive and pain-free for the better quality of life. However, these devices are resource constrained in terms of memory and processing capability. Memory constraint allows these devices to store a limited number of patterns and processing constraint provides delayed response. It is a challenging task to design a robust classification system under above constraints with high accuracy. In this Letter, to resolve this problem, a novel architecture for weightless neural networks (WNNs) has been proposed. It uses variable sized random access memories to optimise the memory usage and a modified binary TRIE data structure for reducing the test time. In addition, a bio-inspired-based genetic algorithm has been employed to improve the accuracy. The proposed architecture is experimented on various disease datasets using its software and hardware realisations. The experimental results prove that the proposed architecture achieves better performance in terms of accuracy, memory saving and test time as compared to standard WNNs. It also outperforms in terms of accuracy as compared to conventional neural network-based classifiers. The proposed architecture is a powerful part of most of the low-power wearable devices for the solution of memory, accuracy and time issues.

Frequent itemset mining is one of the popular techniques used to discover hidden knowledge from large-scale transactional datasets in a wide range of applications. Apriori algorithm is considered as a typical algorithm to find frequent itemsets in market basket analysis. Since its inception, many efforts have been made to enhance the efficiency of the original algorithm. MapReduce model is one of the efficient tools to implement parallel and distributed computing, so that large-scale data set algorithms such as Apriori algorithm can be made efficient in terms of speed up and other related parameters. One of the major drawbacks of the MapReduce model is that it is not suitable for iterative jobs/tasks due to overheads imposed. Now-a-days, Apache Spark is getting huge attention for iterative jobs because of its in-memory processing capabilities. Most of the frequent pattern mining algorithms consider only distinct items in a transaction. For transactional data analysis, multiple occurrences of an item or in other words “quantities” by which a particular item is purchased in the same transaction can be important to derive additional information about frequent itemsets. In this script, we propose a modified version of the Apriori algorithm based on Apache Spark framework that not only mines the frequent itemsets in the input transactional data but also analyzes related quantities of the items for a particular itemset to find the most frequent quantity being purchased for every frequent itemset. Experiments are conducted to gain insight in the form of effectiveness, efficiency, and scalability of the proposed approach.

The Radial Basis Function Neural Networks (RBFNNs) are non-iterative in nature so they are attractive for disease classification. These are four layer networks with input, hidden, output and decision layers. The RBFNNs require single iteration for training the network. On the other side, it suffers from growing hidden layer size on par with training dataset. Though various attempts have been made to solve this issue by clustering the input data. But, in a given dataset estimating the optimal number of clusters is unknown and also it involves more computational time. Hence, to address this problem in this paper, a Particle Swarm Optimization (PSO)-based clustering methodology has been proposed. In this context, we introduce a measure in the objective function of PSO, which allows us to measure the quality of wide range of clusters without prior information. Next, this PSO-based clustering methodology yields a set of High-Performance Cluster Centers (HPCCs). The proposed method experimented on three medical datasets. The experimental results indicate that the proposed method outperforms the competing approaches.

Entity Resolution (ER) is a task for identifying same real world entity. It refers to data object matching or deduplication. It has been a leading research in the field of structure database. Due to its significance, entity resolution continues to be a most important challenge for heterogeneous distributed databases. Several methods have been proposed for the Entity resolution, but they have yielded unsatisfactory results. In this paper, we propose an efficient integrated solution to the entity resolution problem based on Jaccard similarity coefficient. Here we use Markov logic and Jaccard similarity coefficient for providing an efficient solution towards ER problem in heterogeneous distributed databases. The approach that we have implemented gives an overall success rate of about 98 %, thus proving better than the previously implemented algorithms.

As Internet usage is rapidly growing, e-commerce, e-business and corporate world revenue also increases. These areas not only require scalable and consistent databases but also require inter-database transaction support as well. In this paper, we present a scalable architecture along with a distributed algorithm to support atomic transactions across heterogeneous distributed column-oriented databases. Our methodology does not compromise on any assumption on the accuracy of failure territories. Hence, it reveals suited for a class of heterogeneous distributed systems. To achieve such a target, our architectural model exploits an innovative methodology for distributed atomic transactions. Here, we analyse this methodology between different site or node transactions by the analytical performance to make the protocol sufficient.