Boostrapping the Probability Distribution of Peak Electricity Demand
提出用Bootstrap方法估计未来最大电力需求的概率分布，而非仅点预测，并应用于安大略水电公司的实际数据，同时考虑了自变量预测的不确定性。
Demand For effective capacity planning, an electric utility requires an estimate of the probability distribution of future maximum demand, rather than simply a point prediction of expected peak. This paper proposes a method of obtaining this using the bootstrapping technique of B. Efron_(1979) and this is applied to the peak demand of an actual utility, Ontario Hydro. While the technique is constructed from the standard procedure of forecasting a future variable using regression coefficients and known values for the right-hand side variables, it is modified to allow for uncertainty in these independent variable forecasts as well. Copyright 1987 by Economics Department of the University of Pennsylvania and the Osaka University Institute of Social and Economic Research Association.