Non-Dominated Sorting Genetic Algorithm II – Sequential Clustering Classification with Stepwise Dimensionality Reduction Mechanism

• Main Authors: Ying-Ting Tu, 涂穎定
• Other Authors: Chao-Lung Yang
• Format: Others
• Language: en_US
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/uax283

碩士 === 國立臺灣科技大學 === 工業管理系 === 107 === This research aims to improve the performance of the Non-Dominated Sorting Genetic Algorithm – Sequential Clustering Classification (NSGAII-SCC) by integrating a dimensionality reduction technique which is the Principal Component Analysis (PCA). The NSGAII-SCC separates a dataset into two subsets. Next it applies a clustering method to generate labels from a sub-dataset which contains performance related features. Then, the labels generated are used as prediction targets in a classification model to classify data in the second sub-dataset. The sequential clustering and classification (SCC) mechanism is integrated with a feature selection mechanism which is the genetic algorithm (GA). Through the NSGA-II, the compactness of clustering and the prediction accuracy of classification will be simultaneously optimized. Instead of finding the best combination of clustering and classification methods like most studies in the past, this research aims to improve the NSGAII-SCC algorithm performance by integrating a dimensionality reduction technique. The experiment result shows that the proposed method performed better than the original NSGAII-SCC in terms of the solution quality.