Research On Guangxi Electric Power Demand Forecasting Based On ARIMA Model And Regression Analysis

As China’s economy has entered another accelerated growth,the contradiction between energy supply and demand has begun to reappear.As one of China’s important energy sources,electric energy has become more and more important in recent years.Although our country has tried its best to speed up the construction of electric power,the shortage of electricity is still a major problem that cannot be completely solved in the near future.It is estimated that China’s new power generation capacity this year has increased significantly compared to last year,and this trend will continue until next year.This accelerated expansion of power energy will drive the expansion of other related industries,but in fact the actual investment demand is much higher than the investment of the power industry itself.The reasonable speed at which electricity development should be maintained depends largely on the growth in electricity demand.Therefore,the prediction and analysis of power demand has attracted more and more attention from all parties.Then study the internal change law of power consumption and the relationship between power data and its influencing factors,and then make a pre-estimation of power demand.China’s power system has a guiding significance.This paper first introduces the research background of the paper and combs the methods used in the relevant literature on power demand forecasting at home and abroad.Secondly,the research method of this paper is proposed based on the actual situation in Guangxi.This paper proposes a predictive analysis of power demand from both horizontal and vertical perspectives.The basic theories involved are regression analysis and time series prediction.From a horizontal perspective,the current situation of demand for Guangxi Electric Power is analyzed,and the factors affecting the electricity consumption of the whole society are summarized.The experimental data of Guangxi power consumption and its influencing factors from 1990 to 2017 are obtained from the National Bureau of Statistics and the Guangxi Statistical Yearbook.Based on a large amount of actual data,this paper analyzes the changing characteristics of Guangxi’s power consumption and its influencing factors,and expounds the relationship between power demand and Guangxi’s GDP,industrial structure,population factors and other factors.The analysis found that except for the proportion of the tertiary industry’s output value is notsignificant,other factors have a higher degree of correlation.Considering the influence of multicollinearity between independent variables,this paper selects stepwise regression analysis to analyze the power consumption and its influencing factors,establish a regression model and test the regression model and regression coefficient.The test results are significant,indicating that predictive analysis can be performed.And got a better prediction result.From a vertical perspective,according to the time series characteristics of power demand,Eviews software is used to establish a suitable model to verify the stability of the data and determine that the data is not stable.It is necessary to use a differential method to smooth the data and analyze the autocorrelation.And the partial autocorrelation coefficient graph initially identifies a reasonable ARIMA model,and then determines the optimal model based on AIC statistic,SC statistic and HC statistic minimum information criterion,and uses the established optimal model for predictive analysis,and is better.The predicted result.Comparing the prediction results obtained by the two methods,the relative error of the prediction results obtained by the regression analysis prediction method is small.Finally,this paper briefly summarizes the previous analysis and provides comments on future research.