Sample Classification Method for Green Process Evaluation Based on Kernelized Fuzzy C-means Clustering

Abstract

The evaluation sample data of green process have multi-dimensional characteristic, unknown structure, and great dimensional difference. Therefore, the high-quality evaluation sample classification is difficult, when the evaluation is operated by machine learning method. A novel kernelized fuzzy C-means algorithm is proposed in order to achieve the reasonable classification of the samples. Kernelized fuzzy C-means clustering is improved by using the penalty factor; subtraction clustering is applied to enhance the accuracy of clustering; the PBMF index is used as classification condition to obtain the optimal classification number. The experimental results show that this algorithm has good validity and robustness. When the proposed algorithm is applied to classify evaluation samples of green process, the results indicate that the optimal classification of evaluation sample can be implemented effectively