期刊名称：International Journal of Computer Trends and Technology
电子版ISSN：2231-2803
出版年度：2013
卷号：4
期号：5-3
出版社：Seventh Sense Research Group
摘要：An Artificial Neural Network (ANN) does have the capability to provide solutions of various complex problems. The generalization ability of ANN due to the massively parallel processing capability can be utilized to learn the patterns discovered in the data set which can be represented in terms of a set of rules. This rule can be used to find the solution to a classification problem. The learning ability of the ANN is degraded due to the high dimensionality of the datasets. Hence, to minimize this risk we have used Principal Component Analysis (PCA) and Factor Analysis (FA) which provides a feature reduced dataset to the Multi Layer Perceptron (MLP), the classifier used. Again, since the weight matrices are randomly initialized, hence, in this paper we have used Cat Swarm Optimization (CSO) method to update the weight values of the weight matrix. From the experimental evaluation, it was found that using CSO with the MLP classifier provides better classification accuracy as compared to when the classifier is solely used.
