A Performance Degradation Interval Prediction Method Based on Support Vector Machine and Fuzzy Information Granulation

Abstract

To predict the trend and interval of the product performance degradation, a combination approach of fuzzy information granulation (FIG) and support vector machine (SVM) is proposed. Firstly, to make interval prediction of performance degradation and reduce prediction error, the monitoring performance degradation data is divided into several segments in accordance with the actual needs, and the fuzzy information granulation method is used to describe the information of each data segment by the concept of information granule. Then, the support vector machine is applied in the modelling of the fuzzy information granules data. Finally, the proposed FIG–SVM method is applied in degradation assessment of a microwave product, and the result shows that the method is feasible and is effective in improving the modelling precision.