A Hybrid Forecasting Model Based on Empirical Mode Decomposition and the Cuckoo Search Algorithm: A Case Study for Power Load

Figure 10

The comprehensive evaluation of forecasting models from February 2 to March 1. (a) The radar diagram of MAPE by using six different models. (b) Comparison of forecasting MAPE by using the EMD-CSAWNN and ARIMA models. The red line is the polynomial regression line. (c) The comparison of GRA and MAPE by using six different models. (d) The comparison of MAE and RMSE by using six different models. (e) The scatterplot of forecasting versus actual levels by using six different models. The solid line represents the perfect fit: that is, . (f) Evaluation results of six different models. The red font is the best value of every evaluation index; the green font is the worst. (g) The comparison of the box plot by using six different models. The whiskers in the box plot indicate the primary range for the data, in which the lowest data are 1.5 times the interquartile range of the lower quartile and the highest data are 1.5 times the interquartile range of the upper quartile. The outliers, which are not included between the whiskers, are represented by the red crosshair.