Based on large data theory, using data mining technology with the SAS software, we construct a model on the risk prediction of electricityfee recovery using logistic regression.More importantly, we construct separated models for high⁃voltage users, low⁃voltage non⁃family users and low⁃voltage family users based on the market segmentation theory. All the accuracy rates are satis⁃fied, and provide data supporting to cut off the risk of electricity fee recovery and promote the rate of tariff recovery.%在大数据的基础上,通过数据挖掘技术,借助SAS工具,构建了基于逻辑回归的用户电费回收风险预测模型。同时,根据市场细分理论,针对高压用户、低压非居民用户、低压居民用户分别构建了预测模型。预测结果显示:3类模型预测准确率较高,为降低电费回收风险、提升电费回收率提供了数据支撑。
展开▼