摘要
本文采用基于相关性分析的模糊支持向量回归方法,对标准电流互感器的误差试验数据进行预测,并根据预测数据的范围和趋势来判断试验设备的运行状态.其中模糊支持向量回归根据数据的采样时间远近,为离群点赋予不同的权值,并通过遗传算法进行参数寻优,获取最优的预测模型.
Based on the correlation analysis-fuzzy support vector regression （CA-FSVR） method, a prediction model of error test data of the standard current transformer is built; and then the state o{ experiment equipment is thus determined by the scope and the trend of the prediction data. Therein,according to the time dis- tance of the sampling data, different weights are given to the outliers to perform the fuzzy support vector regression method. Parameters of the model are further optimized by the genetic algorithm; finally, the optimal prediction model is obtained.
出处
《三峡大学学报（自然科学版）》
CAS
2016年第6期92-95,共4页
Journal of China Three Gorges University：Natural Sciences
基金
国家自然科学基金项目(61572210
61471177)
关键词
误差试验参数预测
模糊支持向量回归
遗传算法
离群点
error test parameter prediction
fuzzy support vector regression （FSVR）
genetic algorithm
outlier
