GA_SVM

The modern society is prone to many life-threatening diseases which if diagnosis early can be easily controlled. The implementation of a disease diagnostic system has gained popularity over the years. The main aim of this research is to provide a better diagnosis of diabetes. There are already several existing methods, which have been implemented for the diagnosis of diabetes. In this manuscript, firstly, Polynomial Kernel, RBF Kernel, Sigmoid Function Kernel, Linear Kernel SVM used for the classification of PIDD. Secondly GA used as an Attribute selection method and then used Polynomial Kernel, RBF Kernel, Sigmoid Function Kernel, Linear Kernel SVM on that selected attributes of PIDD for classification. So, here compared the results with and without GA in PIDD, and Linear Kernel proved better among all of the noted above classification methods. It directly seems in the paper that GA is removing insignificant features, reducing the cost and computation time and improving the accuracy, ROC of classification. The proposed method can be also used for other kinds of medical diseases.

Speaker Recognition With Normal and Telephonic Assamese Speech Using I-Vector and Learning-Based Classifier

Speaker recognition is the task of identifying a person by his/her unique identification features or behavioural characteristics that are included in the speech uttered by the person. Speaker recognition deals with the identity of the speaker. It is a biometric modality which uses the features of the speaker that is influenced by one's individual behaviour as well as the characteristics of the vocal cord. The issue becomes more complex when regional languages are considered. Here, the authors report the design of a speaker recognition system using normal and telephonic Assamese speech for their case study. In their work, the authors have implemented i-vectors as features to generate an optimal feature set and have used the Feed Forward Neural Network for the recognition purpose which gives a fairly high recognition rate.

Parameter Settings in Particle Swarm Optimization

The issue of parameter setting of an algorithm is one of the most promising areas of research. Particle Swarm Optimization (PSO) is population based method. The performance of PSO is sensitive to the parameter settings. In the literature of evolutionary computation there are two types of parameter settings - parameter tuning and parameter control. Static parameter tuning may lead to poor performance as optimal values of parameters may be different at different stages of run. This leads to parameter control. This chapter has two-fold objectives to provide a comprehensive discussion on parameter settings and on parameter settings of PSO. The objectives are to study parameter tuning and control, to get the insight of PSO and impact of parameters settings for particles of PSO.

A Survey of the Cuckoo Search and Its Applications in Real-World Optimization Problems

The chapter at hand seeks to provide a general survey of the Cuckoo Search Algorithm and its most highlighted variants. The Cuckoo Search Algorithm is a relatively recent nature-inspired population-based meta-heuristic algorithm that is based upon the lifestyle, egg laying, and breeding strategy of some species of cuckoos. In this case, the Lévy flight is used to move the cuckoos within the search space of the optimization problem to solve and obtain a suitable balance between diversification and intensification. As discussed in this chapter, the Cuckoo Search Algorithm has been successfully applied to a wide range of heterogeneous optimization problems found in practical applications over the last few years. Some of the reasons of its relevance are the reduced number of parameters to configure and its ease of implementation.

Optimization Through Nature-Inspired Soft-Computing and Algorithm on ECG Process

In the present research work selection of significant machining parameters depending on nature-inspired algorithm is prepared, during machining alumina-aluminum interpenetrating phase composites through electrochemical grinding process. Here during experimentation control parameters like electrolyte concentration (C), voltage (V), depth of cut (D) and electrolyte flow rate (F) are considered. The response data are initially trained and tested applying Artificial Neural Network. The paradoxical responses like higher material removal rate (MRR), lower surface roughness (Ra), lower overcut (OC) and lower cutting force (Fc) are accomplished individually by employing Cuckoo Search Algorithm. A multi response optimization for all the response parameters is compiled primarily by using Genetic algorithm. Finally, in order to achieve a single set of parametric combination for all the outputs simultaneously fuzzy based Grey Relational Analysis technique is adopted. These nature-driven soft computing techniques corroborates well during the parametric optimization of ECG process.

Quantum-Inspired Computational Intelligence for Economic Emission Dispatch Problem

Economic emission dispatch (EED) problems are one of the most crucial problems in power systems. Growing energy demand, limited reserves of fossil fuel and global warming make this topic into the center of discussion and research. In this chapter, we will discuss the use and scope of different quantum inspired computational intelligence (QCI) methods for solving EED problems. We will evaluate each previously used QCI methods for EED problem and discuss their superiority and credibility against other methods. We will also discuss the potentiality of using other quantum inspired CI methods like quantum bat algorithm (QBA), quantum cuckoo search (QCS), and quantum teaching and learning based optimization (QTLBO) technique for further development in this area.

A New SVM Method for Recognizing Polarity of Sentiments in Twitter

Online social networking platforms, such as Weblogs, micro blogs, and social networks are intensively being utilized daily to express individual's thinking. This permits scientists to collect huge amounts of data and extract significant knowledge regarding the sentiments of a large number of people at a scale that was essentially impractical a couple of years back. Therefore, these days, sentiment analysis has the potential to learn sentiments towards persons, object and occasions. Twitter has increasingly become a significant social networking platform where people post messages of up to 140 characters known as ‘Tweets'. Tweets have become the preferred medium for the marketing sector as users can instantly indicate customer success or indicate public relations disaster far more quickly than a web page or traditional media does. In this paper, we have analyzed twitter data and have predicted positive and negative tweets with high accuracy rate using support vector machine (SVM).

Optimization of Process Parameters Using Soft Computing Techniques

Machining of hard metals and alloys using Conventional machining involves increased demand of time, energy and cost. It causes tool wear resulting in loss of quality of the product. Non-conventional machining, on the other hand produces product with minimum time and at desired level of accuracy. In the present study, EN19 steel was machined using CNC Wire Electrical discharge machining with pre-defined process parameters. Material Removal Rate and Surface roughness were considered as responses for this study. The present optimization problem is single and as well as multi-response. Considering the complexities of this present problem, experimental data were generated and the results were analyzed by using Taguchi, Grey Relational Analysis and Weighted Principal Component Analysis under soft computing approach. Responses variances with the variation of process parameters were thoroughly studied and analyzed; also ‘best optimal values' were identified. The result shows an improvement in responses from mean to optimal values of process parameters.

Swarm Intelligence for Electromagnetic Problem Solving

The chapter will describe the potential of the swarm intelligence and in particular quantum PSO-based algorithm, to solve complicated electromagnetic problems. This task is accomplished through addressing the design and analysis challenges of some key real-world problems. A detailed definition of the conventional PSO and its quantum-inspired version are presented and compared in terms of accuracy and computational burden. Some theoretical discussions concerning the convergence issues and a sensitivity analysis on the parameters influencing the stochastic process are reported.

A Comprehensive Literature Review on Nature-Inspired Soft Computing and Algorithms

Soft Computing techniques are capable of identifying uncertainty in data, determining imprecision of knowledge, and analyzing ill-defined complex problems. The nature of real world problems is generally complex and their common characteristic is uncertainty owing to the multidimensional structure. Analytical models are insufficient in managing all complexity to satisfy the decision makers' expectations. Under this viewpoint, soft computing provides significant flexibility and solution advantages. In this chapter, firstly, the major soft computing methods are classified and summarized. Then a comprehensive review of eight nature inspired – soft computing algorithms which are genetic algorithm, particle swarm algorithm, ant colony algorithms, artificial bee colony, firefly optimization, bat algorithm, cuckoo algorithm, and grey wolf optimizer algorithm are presented and analyzed under some determined subject headings (classification topics) in a detailed way. The survey findings are supported with charts, bar graphs and tables to be more understandable.