A Genetic Algorithm Regression Model for the Mid-long Term of China's Electricity Consumption

Based on the traditional linear regression of the original data, this paper firstly estimates the regression parameters and uses it as the initial domain of the population. Then, according to the timeliness of historical data, the original data of different periods are given corresponding weights. To improve the forecasting accuracy, the genetic algorithm (GA) is introduced to optimize the traditional regression model. This model takes China's 2001-2017 national electricity consumption as a practical example, and compares the above methods with traditional linear regression model and SVM regression model. The results indicate that the forecasting method based on GA-optimized has an acceptable high accuracy.