CAJ | 학술논문

受多因素影响，季度供电量存在增长性及季节波动性的复杂二重趋势。针对现有预测方法中对供电量自身组分及外部影响因素信息利用不足的情况，提出利用行业分类及数据关联度寻优方法选择最优影响参数个数，从而确定模型参数N，构造最优GM（1，N）电量预测模型，提高预测准确度的方法。根据供电企业的用电用户类别，对供电量进行自身组分分析；计算各组分及外部影响因素与供电量的关联度，并对关联度由大到小进行排序；根据不同N的拟合精度确定最优GM（1，N）模型。应用该方法对某供电局的供电量数据进行预测分析表明，该算法具有预测精度好、结果可信度高的特点。
수다인소영향，계도공전량존재증장성급계절파동성적복잡이중추세。침대현유예측방법중대공전량자신조분급외부영향인소신식이용불족적정황，제출이용행업분류급수거관련도심우방법선택최우영향삼수개수，종이학정모형삼수N，구조최우GM（1，N）전량예측모형，제고예측준학도적방법。근거공전기업적용전용호유별，대공전량진행자신조분분석；계산각조분급외부영향인소여공전량적관련도，병대관련도유대도소진행배서；근거불동N적의합정도학정최우GM（1，N）모형。응용해방법대모공전국적공전량수거진행예측분석표명，해산법구유예측정도호、결과가신도고적특점。
Affected by multiple factors, quarter power supply sequence presents the double trends of increasing and fluctuating. Existing forecasting algorithms can’t make full use of all components of power supply data and external influencing factors. In this paper, an optimal GM(1, N) power supply forecasting algorithm is proposed to solve all these problems and improve the forecasting accuracy. And this algorithm is based on industry category and data correlation optimization method to determine the N in the optimal GM(1, N) model. Firstly, according to the power user category of power supply enterprise, the component of power supply data is analyzed. Then the correlation between power supply data and each factor is calculated and sorted. Finally, the GM(1, N) model of different N is estab?lished, and optimal GM(1, N) model is determined according to the fitting precision. The actual power supply data is used to test this forecasting algorithm. The results show that the forecasting algorithm has the characteristics of high accuracy and high reliability.