Authors : B. Renuka Devi, K. Nageswara Rao and S. Pallam Setty
Abstract: This study presents a novel methodology based on Particle Swarm Optimization algorithm to model a new classification system. In the creation of classifier, feature selection frequently used to remove in appropriate and noisy features asto retrieve relevant features. Physically developing a feature set can be very time taking and expensive attempt. PSO is an intelligent search methodology that employs a population of individuals prevailing within a multi-dimensional space. This study employs the correlation between the attributes as the fitness function to Particle Swarm Optimization algorithm. The proposed approach is applied to Clinical Dengue Datasets that retrieve optimal features and the obtained results shows the accuracy and validity of the approach. The proposed methodology is analyzed on dengue data set that is downloaded from http://www.ncbi.nlm.nih.gov/gds. The data set contains 18 attributes of 1275 patients.
B. Renuka Devi, K. Nageswara Rao and S. Pallam Setty, 2016. Towards Better Classification Using Improved Particle Swarm Optimization Algorithm and Decision Tree for Dengue Datasets. International Journal of Soft Computing, 11: 18-25.