Feature Evaluation of Non Dominated Sorting Genetic Algorithm II – Sequential Clustering Classification

• Main Authors: Shih-Chieh Tai, 戴士傑
• Other Authors: Chao-Lung Yang
• Format: Others
• Language: en_US
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/ya3v2e

碩士 === 國立臺灣科技大學 === 工業管理系 === 106 === This research aims to develop a data analysis framework to analyze the data generated by performing clustering and classification sequentially. The clustering and classification procedures are conducted on two different types of data, respectively: one contains the performance measure dataset denoted as Q dataset; the other contains the factors or relevant information regarding the clustering result, denoted as X dataset. Non Dominated Sorting Genetic Algorithm – Sequential Clustering Classification (NSGAII-SCC) with stepwise regression was proposed to investigate the multiple solutions generated from the proposed NSGA which aims to optimize the compactness of clustering and accuracy of classification. In this research, the stepwise regression model and frequency-based similarity matrix were applied to eliminate the redundant features in the datasets and identify the significant features for classification, respectively. The experiment result shows that the proposed methods are able to identify the redundant features and also provide the useful information for the evaluation of clustering and classification results.