Risk warning model of electric power marketing audit work order based on entropy value method

Abstract:
Power marketing audit plays an important role in improving the economic benefits of power enterprises. The current audit work order risk early warning model obtains the early warning results by regression processing a large number of historical data, which is not only inefficient, but also difficult to maintain high early warning accuracy. In order to improve the above defects, the risk early warning model of power marketing audit work order based on entropy method will be studied and constructed. After mining work order data by decision tree algorithm, the association between work order businesses is calculated. According to the basic principle of entropy method, two risk indexes of emergency degree and thermal degree are established from the dimensions of time and region. The BP neural network is improved to get the final risk early warning model. The case study shows that the early warning model is not only efficient, but also has an error of 0.12%, which has higher practical application value.