A Integrated IFCM-MPSO-SVM Model for Forecasting Equipment Support Capability

Yong Shuai ； Tai-liang Song ； Jian-ping Wang ； Hong Shen ； Wen-bin Zhan

《電腦學刊》 28卷1期 (2017/02) Pp. 233-245

https://doi.org/10.3966/199115592017122801018

摘要

For the sake of improving the accuracy for forecasting equipment support capability, aiming at the problems in support vector machine forecast model, this paper improved fuzzy Cmeans clustering algorithm about outliers operation and optimization of distance in the clusters and among the clusters firstly. Then this method was used to optimize the input feature sets and reduce the redundancy and excess of the training sample sets. Furthermore, confirmed the Radial Basis Function by comparing the character of the kernel functions. At the same time, modified the particle swarm optimization algorithm about the particle speed, location and the inertia weight value to increase the diversity of particle swarm and avoided the convergence of searching, and this method is used to optimize the SVM parameters and built the forecast model. Finally the example showed the forecast index was objective and the modified forecasting model was accurate.

關鍵字

forecasting ； FP-growth association rule ； fuzzy Bayesian network ； fuzzy C-mean clustering ； modified particle swarm optimization ； support vector machine