Haider Banka

A Hamming distance based binary particle swarm optimization (HDBPSO) algorithm for high dimensional feature selection, classification and validation

Gene expression data typically contain fewer samples (as each experiment is costly) and thousands of expression values (or features) captured by automatic robotic devices. Feature selection is one of the important and challenging tasks for this kind of data where many traditional methods failed and evolutionary based methods were succeeded. In this study, the initial datasets are preprocessed using a quartile based fast heuristic technique to reduce the crude domain features which are less relevant in categorizing the samples of either group. Hamming distance is introduced as a proximity measure to update the velocity of particle(s) in binary PSO framework to select the important feature subsets. The experimental results on three benchmark datasets vis-a-vis colon cancer, defused B-cell lymphoma and leukemia data are evaluated by means of classification accuracies and validity indices as well. Detailed comparative studies are also made to show the superiority and effectiveness of the proposed method. The present study clearly reveals that by choosing proper preprocessing method, fine tuned by HDBPSO with Hamming distance as a proximity measure, it is possible to find important feature subsets in gene expression data with better and competitive performances.

A particle swarm optimization based energy efficient cluster head selection algorithm for wireless sensor networks

Clustering has been proven to be one of the most efficient techniques for saving energy of wireless sensor networks (WSNs). However, in a hierarchical cluster based WSN, cluster heads (CHs) consume more energy due to extra overload for receiving and aggregating the data from their member sensor nodes and transmitting the aggregated data to the base station. Therefore, the proper selection of CHs plays vital role to conserve the energy of sensor nodes for prolonging the lifetime of WSNs. In this paper, we propose an energy efficient cluster head selection algorithm which is based on particle swarm optimization (PSO) called PSO-ECHS. The algorithm is developed with an efficient scheme of particle encoding and fitness function. For the energy efficiency of the proposed PSO approach, we consider various parameters such as intra-cluster distance, sink distance and residual energy of sensor nodes. We also present cluster formation in which non-cluster head sensor nodes join their CHs based on derived weight function. The algorithm is tested extensively on various scenarios of WSNs, varying number of sensor nodes and the CHs. The results are compared with some existing algorithms to demonstrate the superiority of the proposed algorithm.

Local pattern transformation based feature extraction techniques for classification of epileptic EEG signals

Abstract Background and objective According to the World Health Organization (WHO) epilepsy affects approximately 45–50 million people. Electroencephalogram (EEG) records the neurological activity in the brain and it is used to identify epilepsy. Visual inspection of EEG signals is a time-consuming process and it may lead to human error. Feature extraction and classification are two main steps that are required to build an automated epilepsy detection framework. Feature extraction reduces the dimensions of the input signal by retaining informative features and the classifier assigns a proper class label to the extracted feature vector. Our aim is to present effective feature extraction techniques for automated epileptic EEG signal classification. Methods In this study, two effective feature extraction techniques (Local Neighbor Descriptive Pattern [LNDP] and One-dimensional Local Gradient Pattern [1D-LGP]) have been introduced to classify epileptic EEG signals. The classification between epileptic seizure and non-seizure signals is performed using different machine learning classifiers. The benchmark epilepsy EEG dataset provided by the University of Bonn is used in this research. The classification performance is evaluated using 10-fold cross validation. The classifiers used are the Nearest Neighbor (NN), Support Vector Machine (SVM), Decision Tree (DT) and Artificial Neural Network (ANN). The experiments have been repeated for 50 times. Results LNDP and 1D-LGP feature extraction techniques with ANN classifier achieved the average classification accuracy of 99.82% and 99.80%, respectively, for the classification between normal and epileptic EEG signals. Eight different experimental cases were tested. The classification results were better than those of some existing methods. Conclusions This study suggests that LNDP and 1D-LGP could be effective feature extraction techniques for the classification of epileptic EEG signals.

Energy efficient clustering algorithms for wireless sensor networks: novel chemical reaction optimization approach

Clustering has been accepted as one of the most efficient techniques for conserving energy of wireless sensor networks (WSNs). However, in a two-tiered cluster based WSN, cluster heads (CHs) consume more energy due to extra overload for receiving data from their member sensor nodes, aggregating them and transmitting that data to the base station (BS). Therefore, proper selection of CHs and optimal formation of clusters play a crucial role to conserve the energy of sensor nodes for prolonging the lifetime of WSNs. In this paper, we propose an energy efficient CH selection and energy balanced cluster formation algorithms, which are based on novel chemical reaction optimization technique (nCRO), we jointly called these algorithms as novel CRO based energy efficient clustering algorithms (nCRO-ECA). These algorithms are developed with efficient schemes of molecular structure encoding and potential energy functions. For the energy efficiency, we consider various parameters such as intra-cluster distance, sink distance and residual energy of sensor nodes in the CH selection phase. In the cluster formation phase, we consider various distance and energy parameters. The algorithm is tested extensively on various scenarios of WSNs by varying number of sensor nodes and CHs. The results are compared with original CRO based algorithm, namely CRO-ECA and some existing algorithms to demonstrate the superiority of the proposed algorithm in terms of energy consumption, network lifetime, packets received by the BS and convergence rate.

PSO-Based Multiple-sink Placement Algorithm for Protracting the Lifetime of Wireless Sensor Networks

Optimal deployment of multiple sinks has been proven to be one of the energy efficient techniques for prolonging the lifetime of wireless sensor networks (WSNs). In this paper, we propose a particle swarm optimization (PSO) based algorithm called PSO-MSPA for placement of multiple-sink in WSNs. The algorithm is developed with an efficient scheme of particle encoding and novel fitness function. For the energy efficiency of the PSO-MSPA, we consider various parameters such as Euclidian distance and hop count from the gateways to the sinks. The algorithm is tested extensively on various scenarios of WSNs by varying number of gateways and sensor nodes and the results are analyzed to show the efficacy of the proposed algorithm.

BERA: a biogeography-based energy saving routing architecture for wireless sensor networks

Biogeography-based optimization (BBO) is a relatively new paradigm for optimization which is yet to be explored to solve complex optimization problems to prove its full potential. In wireless sensor networks (WSNs), optimal cluster head selection and routing are two well-known optimization problems. Researchers often use hierarchal cluster-based routing, in which power consumption of cluster heads (CHs) is very high due to its extra functionalities such as receiving and aggregating the data from its member sensor nodes and transmitting the aggregated data to the base station (BS). Therefore, proper care should be taken while selecting the CHs to enhance the life of the network. After formation of the clusters, data to be routed to the BS in inter-cluster fashion for further enhancing the life of WSNs. In this paper, a biogeography-based energy saving routing architecture (BERA) is proposed for CH selection and routing. The biogeography-based CH selection algorithm is proposed with an efficient encoding scheme of a habitat and by formulating a novel fitness function that uses residual energy and distance as its metrics. The BBO-based routing algorithm is also proposed. The efficient encoding scheme of a habitat is developed, and its fitness function considers the node degree in addition to residual energy and distance. To exhibit the performance of BERA, it is extensively tested with some existing routing algorithms such as DHCR, Hybrid routing, EADC and some bio-inspired algorithms, namely GA and PSO. Simulation results confirm the superiority/competitiveness of the proposed algorithm over existing techniques.

A Gravitational Search Algorithm for Energy Efficient Multi-sink Placement in Wireless Sensor Networks

Optimal placement of multi-sink has been accepted an energy efficient approaches of extending the life of wireless sensor networks (WSNs). In this paper, a Gravitational Search Algorithm (GSA) based approach called GSA-MSP (Gravitational Search Algorithm based Multi-Sink Placement) for multi-sink placement for sensor network has been proposed. The algorithm has been designed with proper encoding scheme and a new fitness function. We consider the energy, Euclidian distance from the gateways to the sinks, and data rate of gateways are as parameters for the efficient design of GSA-MSP. The GSA-MSP has been tested vigorously over a varying number of sensors, gateways and sinks on various scenarios of WSNs. To show the efficacy of the GSA-MSP has been compared with some existing algorithms.

Feature Selection, Classification and Rule Generation Using Rough Sets

In this chapter we present two applications using rough sets. The first application deals with an evolutionary-rough feature selection algorithm for classifying microarray gene expression patterns. Since the data typically consists of a large number of redundant features, an initial redundancy reduction of the attributes is done to enable faster convergence. Rough set theory is employed to generate reducts, which represent the minimal sets of non-redundant features capable of discerning between all objects, in a multi-objective framework. The effectiveness of the algorithm is demonstrated on three cancer datasets. The second application is concerned with the dependencies among the attributes, their significance, and evaluation performed using intelligent data analysis tool. The predictive model, based on rough set approach generates fewer number of decision rules. It is found that without considering lag attribute for stock price movement decision table, the number of decision rules generated reduces significantly considering to the lag attribute taking into consideration. Ignoring the lag attribute does not affect the degree of dependency while the rest of the conditional attributes are same. The results are also compared with the neural network based algorithm. Rough confusion matrix is used to evaluate the predicted classification performances.

FARW: Firefly algorithm for Routing in wireless sensor networks

Firefly algorithm is a new nature inspired algorithm for optimization that needs to be explored further to solve a number of problems to show its real potential. Routing is a well known techniques for prolonging the life of the wireless sensor networks. The current work focuses on solving aforementioned related problem using firefly algorithm. The Routing algorithm is developed with novel fitness function based on residual energy, node degree and distance. The proposed algorithms are extensively tested on various scenarios to show its performance and compared with conventional algorithms such as EADC, DHCR and Hybrid Routing. Experimental results depicts that proposed algorithms performs better as compared with some existing one.

CRHS: clustering and routing in wireless sensor networks using harmony search algorithm

In wireless sensor networks, cluster head selection and routing are two well-known optimization problems associated with high computational complexity. Harmony search algorithm (HSA) is one of the metaheuristics, used to solve a wide range of NP-Hard problems. In this paper, first we propose an HSA-based cluster head (CH) selection algorithm by devising a fitness function with energy, distance and node degree as parameters. Next, we derived a potential function for the assignment of non-CH nodes to the CHs. Finally, an HSA-based routing algorithm is also proposed using the same parameters, i.e., energy, distance and node degree in the derivation of the fitness function. Three test cases have been considered in this study for performance evaluation. The proposed algorithm has been tested with some of the existing related techniques. Simulation results depict that the proposed algorithm (CRHS) shows superior performance over the existing techniques.