Suitable Features Selection for the HMLP Network Using Circle Segments Method

This paper integrates the circle-segments as a feature selection method with the Hybrid MultiLayer Perceptron (HMLP) neural network. The circle-segments method is used to provide visual correlation between the input-output data samples to users, and to allow users to eliminate insignificant inputs from the input data set, while the HMLP network is employed to tackle pattern classification tasks. The effectiveness of the HMLP-circle segments system is evaluated using the wine and aggregate problems. The results are compared with those from the HMLP coupled with the Principal Component Analysis (PCA) as well as HMLP without any feature selection method. It is found that the accuracy of the HMLP-circle segments system is better than, HMLP coupled with the PCA and HMLP without any feature selection method.