CAJ | 학술논문

为有效预测呈指数增长的年度用电量，本文选用了指数函数法、GM（1，1）灰色模型和多元线性回归模型来对年度用电量进行拟合外推预测。通过运用这些方法预测北京年度用电量和对预测结果进行比较分析得到，GM（1，1）灰色模型的实际预测结果往往偏大，指数函数法预测结果可能偏小，而多元线性回归模型虽然考虑了其他因素影响但在实际预测指数增长负荷时参考意义较小。N此，最终选择合适的权重加权组合以上模型得到最终预测值，预测精度较高。
위유효예측정지수증장적년도용전량，본문선용료지수함수법、GM（1，1）회색모형화다원선성회귀모형래대년도용전량진행의합외추예측。통과운용저사방법예측북경년도용전량화대예측결과진행비교분석득도，GM（1，1）회색모형적실제예측결과왕왕편대，지수함수법예측결과가능편소，이다원선성회귀모형수연고필료기타인소영향단재실제예측지수증장부하시삼고의의교소。N차，최종선택합괄적권중가권조합이상모형득도최종예측치，예측정도교고。
To forecast the annual electricity consumption of exponent increase trend, extrapolation forecasting mathematical Junctions, such as ex- ponential function, GM （1,1） gray model and multiple linear regression model, were adopted to directly fit the historical loads and forecast the fu- ture annual loads. Through using these new functions to fit the historical data of Beijing and giving a comparative analysis of the forecasting re- suits, it can be obtained that, the predicted value of GM （1,1） gray model is bigger than the actual loads, and the predicted value of exponential function is smaller than the actual loads. Considering other factors that affect the annual load, multiple linear regression model provides less refer- ence value to actual annual load forecasting of exponential growth. Ultimately, the author gives the final combination predicted results based on proper weights. The mean relative percentage error of the combination forecasting method is less than 2 ~.