基于数据挖掘的电信故障分类及回归预测

The Classification and Regression Prediction of Telecom Fault Based on Data Mining

基于数据挖掘技术,针对电信故障海量数据特点,合理选择属性值和标签值,运用交叉验证、网格划分、遗传算法和粒子群算法进行参数寻优,运用支持向量机SVM理论,建立电信故障分类模型和预测模型。通过仿真分析,并且与电信故障实际数据对比,表明该分类模型和预测模型的精度高,误差小,为今后控制电信故障,改善网络运行质量提供理论依据和数据支持。

Based on data mining through the detailed analysis of telecom fault massive data features,reasonably choosing data value and label value,using cross validation,grid,GA and PSO to parameter optimization,along with the theory of SVM to establish the telecom fault classification model and prediction model.Through comparative analysis of prediction data and actual measurement data,it shows that this classification model and prediction model has a high level in telecom fault forecasting with low errors.All of these can provide data and theoretical support for telecom fault controlling and improving the quality of network in future.