《江苏科技信息》杂志社

Jiangsu Science & Technology Information Magazine Agency

关键词: 数据挖掘, 多元线性回归, 分时段

Abstract: The number prediction of accepted fault repairs is an important part of grid operation and power supply services. Predicting the number of accepted fault repairs scientifically and accurately is good for the healthy development of the power industry, and the improvement of the people's living standards. This paper uses the multiple linear regression algorithms as the data mining method, building the time-segment prediction model about the number of accepted fault repairs based on the maximum temperature, the maximum wind, the rainfall amount, the humidity, and the historical data, and testing the multiple linear regression model's validity by comparing various prediction algorithm. The model provides the scientific basis for the number prediction of accepted fault repairs; thereby ensures the safety and stability of the power grid and provides high quality services for the power users.