Variable Optimization in Cervical Cancer Data Using Particle Swarm Optimization

摘要

Samples of data may consist of numerous attributes and variables which are irrelevant and redundant. Some of those attributes may not be of any vital use in classification and the irrelevant attributes can decrease the efficiency. Thus, the feature reduction process can be considered as a problem in machine learning which selects less quantity of vital attributes to obtain higher accuracy rate. This process minimizes the attributes count by eliminating less relevant and noisy samples from the data set to achieve better classification accuracy. This work uses particle swarm optimization (PSO) search algorithm for feature reduction in cervical cancer data set. The experimental result shows that the irrelevant features are removed and only 17 useful features are selected, out of which 36 in the cervical cancer data set.