Research on prediction of power market credit system based on linear model and Improved BP neural network

Abstract

With the continuous economic growth, the number of power customers has increased significantly, and consumers in the field of power marketing will inevitably have a credit crisis. In order to reduce the business risk of relevant departments and improve the risk prediction ability of the system, this paper evaluates and reviews the user credit system. In this paper, the basic structure of BP neural network is described firstly, and then the traditional BP neural network model is optimized after analyzing its algorithm flow. Based on this point, this paper analyzes the characteristics of customers in the energy and electricity market in the research area, and referring to local experts who have been engaged in power sales for many years, this paper puts forward a new set of directly scored power system load forecasting index system and algorithm improvement scheme, and discusses the evaluation of power market credit rating based on the credit evaluation suggestions of power customers. After establishing the judgment criteria, in this paper, the power load data of the target area is studied by empirical analysis method, and selects three different customers from the production area, commercial and residential areas and residential areas as cases to analyze the determination of their credit rating, and then discusses the results of regional power load forecasting. Finally, this paper puts forward a kind of power management method based on the user's credit rating, and in order to complete the modernization transformation of power management system and promote the market development. In this paper, after improving and optimizing the traditional BP neural network, it is applied to the power market to predict the target user credit system, so as to achieve the improvement of forecasting ability.

Carbon Emission Influencing Factors and Scenario Prediction for Construction Industry in Beijing–Tianjin–Hebei

“…In this paper, the grey model (1, 1) and autoregressive integrated moving average models (abbreviated as GM (1, 1), ARIMA) with simple operation and high accuracy are selected to fit the prediction of direct carbon emissions from the construction industry in Beijing-Tianjin-Hebei to avoid the unscientific evaluation of model accuracy by a single error indicator. Four performance indicators, including mean square error (MSE), root mean square error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE), were selected to evaluate the prediction model accuracy, and the calculation formula was as follows [48]:…”