Skip to main content

Über dieses Buch

This edited volume on computational intelligence algorithms-based applications includes work presented at the International Conference on Computational Intelligence, Communications, and Business Analytics (CICBA 2017). It provides the latest research findings on the significance of computational intelligence and related application areas. It also introduces various computation platforms involving evolutionary algorithms, fuzzy logic, swarm intelligence, artificial neural networks and several other tools for solving real-world problems. It also discusses various tools that are hybrids of more than one solution framework, highlighting the theoretical aspects as well as various real-world applications.



Linear Programming-Based TOPSIS Method for Solving MADM Problems with Three Parameter IVIFNs

The aim of this paper is to develop a TOPSIS approach using fractional programming techniques for effective modelling of real-life multiattribute decision-making (MADM) problems in interval-valued intuitionistic fuzzy (IVIF) settings by considering hesitancy degree as a dimension together with membership and non-membership degrees. In three-parameter characterizations of intuitionistic fuzzy (IF) sets, a weighted absolute distance between two IF sets with respect to IF weights is defined and employed in TOPSIS to formulate intervals of relative closeness coefficients (RCCs). The lower and upper bounds of the intervals of RCCs are given by a pair of nonlinear fractional programming models which are further transformed into two auxiliary linear programming models using mathematical methods and fractional programming technique. A simpler technique is also proposed for estimating the optimal degrees as performance values of alternatives from the possibility degree matrix generated by pairwise comparisons of RCC intervals. The validity and effectiveness of the proposed approach are demonstrated through two numerical examples.
Samir Kumar, Animesh Biswas

A Comparative Study of Bio-inspired Algorithms for Medical Image Registration

The challenge of determining optimal transformation parameters for image registration has been treated traditionally as a multidimensional optimization problem. Non-rigid registration of medical images has been approached in this article using the particle swarm optimization algorithm, dragonfly algorithm, and the artificial bee colony algorithm. Brief introductions to these algorithms have been presented. Results of MATLAB simulations of medical image registration approached through these algorithms have been analyzed. The simulation shows that the dragonfly algorithm results in higher quality image registration, but takes longer to converge. The trade-off issue between the quality of registration and the computing time has been brought forward. This has a strong impact on the choice of the most suitable algorithm for medical applications, such as monitoring of tumor progression.
D. R. Sarvamangala, Raghavendra V. Kulkarni

Different Length Genetic Algorithm-Based Clustering of Indian Stocks for Portfolio Optimization

In this chapter, we propose a model for portfolio construction using different length genetic algorithm (GA)-based clustering of Indian stocks. First, stocks of different companies, chosen from different industries, are classified based on their returns per unit of risk using an unsupervised method of different length genetic algorithm. Then, the centroids of the algorithm are again classified by the same algorithm. So vertical clustering (clustering of stocks by returns per unit of risk for each day) followed by horizontal clustering (clustering of the centroids over time) eventually produces a limited number of stocks. The Markowitz model is applied to determine the weights of the stocks in the portfolio. The results are also compared with some well-known existing algorithms. They indicate that the proposed GA-based clustering algorithm outperforms all the other algorithms.
Somnath Mukhopadhyay, Tamal Datta Chaudhuri

An Evolutionary Matrix Factorization Approach for Missing Value Prediction

Sparseness of data is a common problem in many fields such as data mining and pattern recognition. During the last decade, collecting opinions from people has been established to be an useful tool for solving different real-life problems. In crowdsourcing systems, prediction based on very few observations leads to complete disregard for the inherent latent features of the crowd workers corresponding to the items. Similarly in bioinformatics, sparsity has a major negative impact in finding relevant gene from gene expression data. Although this problem is being studied over the last decade, there are some benefits and pitfalls of the different proposed approaches. In this article, we have proposed a genetic algorithm-based matrix factorization technique to estimate the missing entries in the rating matrix of recommender systems. We have created four synthetic datasets and used two real-life gene expression datasets to show the efficacy of the proposed method in comparison with the other state-of-the-art approaches.
Sujoy Chatterjee, Anirban Mukhopadhyay

Differential Evolution in PFCM Clustering for Energy Efficient Cooperative Spectrum Sensing

Cooperative spectrum sensing (CSS) in cognitive radio network (CRN) is highly recommended to avoid the interference from secondary users (SUs) to primary user (PU). Several studies report that clustering-based CSS technique improves the system performance, among them fuzzy c-means (FCM) clustering algorithm is widely explored. However, it is observed that FCM generates an improper clustering of sensing information at low signal-to-noise ratio (SNR) due to inseparable nature of energy data set. To address this problem, the present chapter describes a work that investigates the scope of possibilistic fuzzy c-means (PFCM) algorithm on energy detection-based CSS. PFCM integrates the possibilistic information and fuzzy membership values of input data in the clustering process to segregate the indistinguishable energy data into the respective clusters. Differential evolution (DE) algorithm is applied with PFCM to maximize the probability of PU detection (\(P_D\)) under the constraint of a target false alarm probability (\(P_{fa}\)). The present work also evaluates the required power consumption during CSS by SUs. The proposed technique improves \(P_D\) by \(\sim \!12.53\%\) and decreases average energy consumption by \(\sim \!5.34\%\) over the existing work.
Anal Paul, Santi P. Maity

Feature Selection for Handwritten Word Recognition Using Memetic Algorithm

Nowadays, feature selection is considered as a de facto standard in the field of pattern recognition where high-dimensional feature attributes are used. The main purpose of any feature selection algorithm is to reduce the dimensionality of the input feature vector while improving the classification ability. Here, a Memetic Algorithm (MA)-based wrapper–filter feature selection method is applied for the recognition of handwritten word images in segmentation-free approach. In this context, two state-of-the-art feature vectors describing texture and shape of the word images, respectively, are considered for feature dimension reduction. Experimentation is conducted on handwritten Bangla word samples comprising 50 popular city names of West Bengal, a state of India. Final results confirm that for the said recognition problem, subset of features selected by MA produces increased recognition accuracy than the individual feature vector or their combination when applied entirely.
Manosij Ghosh, Samir Malakar, Showmik Bhowmik, Ram Sarkar, Mita Nasipuri

A Column-Wise Distance-Based Approach for Clustering of Gene Expression Data with Detection of Functionally Inactive Genes and Noise

Due to uncertainty and inherent noise present in gene expression data, clustering of the data is a challenging task. The common assumption of many clustering algorithms is that each gene belongs to a cluster. However, few genes are functionally inactive, i.e. not participate in any biological process during experimental conditions and should be segregated from clusters. Based on this observation, a clustering method is proposed in this article that clusters co-expressed genes and segregates functionally inactive genes and noise. The proposed method formed a cluster if the difference in expression levels of genes with a specified gene is less than a threshold t in each experimental condition; otherwise, the specified gene is marked as functionally inactive or noise. The proposed method is applied on 10 yeast gene expression data, and the result shows that it performs well over existing one.
Girish Chandra, Sudhakar Tripathi

Detection of Moving Objects in Video Using Block-Based Approach

In this paper, an efficient technique has been proposed to detect moving objects in the video under dynamic as well as static background condition. The proposed method consists block-based background modelling, current frame updating, block processing of updated current frame and elimination of background using bin histogram approach. Next, enhanced foreground objects are obtained in the post-processing stage using morphological operations. The proposed approach effectively minimizes the effect of dynamic background to extract the foreground information. We have applied our proposed technique on Change Detection CDW-2012 dataset and compared the results with the other state-of-the-art methods. The experimental results prove the efficiency of the proposed approach compared to the other state-of-the-art methods in terms of different evaluation metrics.
Amlan Raychaudhuri, Satyabrata Maity, Amlan Chakrabarti, Debotosh Bhattacharjee


Weitere Informationen

Premium Partner