電力在人們的生活中占很大的一部分,不論是民生日用品或是工廠生產機具都與之息息相關,因此如何有效的用電成了台電公司必須面對的課題,工作天數是判斷用電戶工作型態的基本方法,透過工作天數的判斷,可得知公司不需用電的時段,台電公司便可依據此數據減少對該公司的供電,在本研究中以自我相關函數對用戶的用電資料進行運算,並搭配迴歸分析的方式,增加可計算之用戶數量並節省收集資料的時間,關聯分析的部分則是針對用電戶屬性如契約別、用電別、區處別、行業別、基本電費、流動電費、經常契約容量、最高用電量、最低用電量、平均用電量及工作天數等11個屬性進行相關係數的計算,從中選出高度相關之屬性,以文獻及其他分析方式加以輔助證明此關聯分析之結果。
The power is the most important in part of our life, no matter people’s articles for daily use or factory to produce machines. So, how to use the power effectively is the problem that Taiwan Power Company has to face. Work days are the basic method to judge what work styles the power customers are, through the work days of power customers the company can know the time of power that has not been used, the Taiwan Power Company can reduce the power of the company according to the data. In this study, computing the power customers’ electricity data with Autocorrelation Function (ACF) and method of regression analysis to increase the number of calculable power customers, and save time of collecting information. The part of correlativity analysis is for power customers attributes, for example: ContractID, ElectricityID, AreaID, BusinessID, Basic electricity bill, Flow electricity bill, Regular contract capacity, The max electricity consumption, The min electricity consumption, The mean electricity consumption, Work days and so on. Using 11 attributes to calculate the correlation coefficient, and select the attributes which are the high correlation of them, using literature and other analytical methods to prove the results of the correlation analysis.