标题：Evaluation of the Application of a Forecasting Model Using the Classic Least Squares Method in Forecasting Daily 15-Minute Peak Power Demand in the National Power System
摘要：The paper discusses the results of historical (ex post) forecasting of the 15-minute peak demand for power in the National Power System (NPS).The study employed the classic least squares method based on the autoregressive nature of the analysed quantity (without explanatory vari- ables) only.Simulation tests in the expired mode (with known value of the forecast variable) for the following day included analyses for 2nd and 3rd order polynomials,for the number of values delayed by from 2 to 16 preceding days,and the purpose of the study was to select their most favourable combinations.The analysed time series covered the periods of 13 years and 5 years divided into weekdays.The forecast results (ex post and ex ante) were compared with naive fore- casts (for weekly delays) assuming the assignment of earlier values to the forecast quantities.Effectiveness of the most favourable expired prediction for the 3rd order polynomial and the 15-day delay obtained by the classic least squares method was lower than for the naive forecasts.
关键词：forecasting;electric power demand;national power system load;classic least squares error method
