摘要
利用基于时间序列的ARIMA算法对用户的历史用电行为、交费行为、资金到账规律进行研究分析、建模,并预测售电收入日现金流。基于研究结果,预测了未来两个月的月度售电收入资金,月均偏差率在2%以内,当月25日预测当月月度售电收入资金月均偏差率在1%以内;预测了未来两个月的日售电收入到账金额,并分析预测当月日累计收入偏差率在2%以内的日期规律,当月25日至月底预测当月日累计收入资金日均偏差率在1%以内。
This paper uses time series,regression,correlation analysis,clustering and other big data algorithms to research,analyze and model the user's historical electricity usage behavior,payment behavior,and capital arrival rules,and predict the daily cash flow of electricity sales revenue.Based on the research results,the monthly sales revenue of the next two months is predicted,and the monthly average deviation rate is within 2%.The monthly average deviation rate of monthly sales revenue is predicted to be within 1%of the month.
出处
《工业控制计算机》
2018年第12期58-59,62,共3页
Industrial Control Computer
关键词
月度售电收入
日售电收入
现金流
预测模型
大数据算法
monthly sales revenue
daily sales revenue
cash flow
forecasting model
big data algorithm
