Dharmpal Singh

A comparative study on the performance of fuzzy logic, Bayesian logic and neural network towards decision-making

Soft computing models play an important role in the field of recognition, classification, data prediction, etc., and also in various application fields towards decision-making. Soft computing models include fuzzy logic, neural, network, genetic algorithm, particle swarm optimisation, tabu search, harmonie search, clustering, etc. The performance of a particular soft computing model can be ascertained using a particular dataset for the purpose of decision-making. Here, an effort has been made to make a comparison on the performance of fuzzy logic, Bayesian logic and neural network. The model with minimum error has been given preference for selection towards decision-making of information. The same method has been cross-checked based on the residual analysis to verify the earlier proposed observation. The said models have also been cross-checked based on other dataset. Under neural network, perceptron neural network model has been used.

An effort to select a preferable metaheuristic model for knowledge discovery in data mining

Swarm intelligence computation is becoming an important topic in the field of decision making, and is successfully applied in many fields, which is indicating a fairly great potential for development. Recently, much research work has been carried out by swarm intelligence in different fields viz. sensor network, robotic management, microwave filter design, data mining applications etc. In this paper, an effort has been made to make a comparison on the performance of fuzzy logic with different membership function, artificial neural network and swarm intelligence models. The models with minimum errors have been given preference for selection towards the decision making of information. The same methods can also be cross-checked by residual analysis and a new unknown data set to verify the earlier proposed observation. Initially the statistical techniques have been applied on the data set to select the preferable models but it has been observed that statistical technique errors are large. Therefore, under the fuzzy logic, different membership functions have been used to select the preferable membership function for fuzzy logic for the available data. In this paper further effort has also been made to change the expression of the computing fitness function of the path in artificial bee colony to achieve the best result.

An Improved Bio-inspired BAT Algorithm for Optimization

Metaheuristic algorithms are used today to solve many optimization-related problems. The firefly algorithm, particle swarm optimization, harmony search and BAT algorithm are used as metaheuristic searched algorithms to find the optimized solution of the problem domain. The BAT algorithm was developed by using the unique characteristics of BAT which used the advanced capability of echolocation to move in the dark, avoid obstacles or barrier and also find its food or pray. The main aim of this paper is to represent and translate the behaviour of BAT algorithm in the form of improved BAT algorithm. The paper also describes the implication, advantage, disadvantage and application of BAT algorithm in different areas used by diversified authors.