A Multi-Objective Particle Swarm Optimization For Feature Selection

ISBN-13: 9786139864638 / Angielski / Miękka / 2018 / 60 str.

Y. Mohana Roopa; D.Kishore Babu

Opis

Feature selection is very important task in classification. Number of features is available for classification but not all of them are useful. Irrelevant and redundant features may even reduce the performance. There are two types of feature selection approaches. They are wrapper and filter approaches. Their main difference is that wrappers use a classification algorithm when searching the goodness of the features during the feature selection process while filters are independent of any classification algorithm. The goal of Feature selection is to choose a small number of relevant features to achieve similar or even better classification performance than using all features. Existing feature selection algorithms treat the task as a single objective problem. The proposed system can treat as a multi objective problem. It has two objectives. They are maximizing the classification performance and minimizing the number of features. The proposed system is PSO-based multi-objective feature selection algorithm. The algorithm (NSPSOFS) introduces the task is to generate a Pareto front of non dominated solutions idea of non dominated sorting into PSO to address feature selection problems.

O autorze

Roopa, Y. Mohana
Dr. Y Mohana Roopa is the Professor of Computer science and Engineering. She is also Dean of Continuing Education and Internal Quality Assurance at Institute of Aeronautical Engineering. She has 18 years of experience in teaching, R & D. She has published 40 papers in international and national journals and conferences.