Research Article

Electric Load Forecast Using Combined Models with HP Filter-SARIMA and ARMAX Optimized by Regression Analysis Algorithm

Table 7

Predictions of five models.

Obs.Actual loadActual temperatureHP Filter-SARIMA-Revised ARMAXSARIMAARIMAGrey system MLP neural network
ForecastError (%)ForecastError (%)ForecastError (%)ForecastError (%)ForecastError (%)

16th Jul.458535.04570−0.3272%47483.5551%4287−6.4995%45870.0436%4428−3.4242%
17th Jul.473339.44699−0.7184%49905.4300%4210−11.0501%4457−5.8314%4558−3.6974%
18th Jul.482439.94799−0.5182%51586.9237%4320−10.4478%4657−3.4619%4721−2.1352%
19th Jul.501241.94989−0.4589%552010.1357%4580−8.6193%4851−3.2123%4518−9.8563%
20th Jul.429331.943080.3494%482412.3690%4127−3.8668%44273.1214%4212−1.8868%
21st Jul.447039.045100.8949%47506.2640%4215−5.7047%46393.7808%44980.6264%
22nd Jul.458940.84588−0.0218%46501.3293%4218−8.0846%4418−3.7263%4357−5.0556%
23rd Jul.463033.34599−0.6695%4598−0.6911%4324−6.6091%4417−4.6004%4500−2.8078%
24th Jul.449735.94485−0.2668%4325−3.8248%4129−8.1832%4213−6.3153%4367−2.8908%
25th Jul.411629.341200.0972%3998−2.8669%3950−4.0330%3854−6.3654%4007−2.6482%
26th Jul.383526.43827−0.2086%3560−7.1708%3648−4.8761%3712−3.2073%3714−3.1551%
27th Jul.372627.13697−0.7783%3589−3.6769%3457−7.2195%3320−10.8964%3421−8.1857%
28th Jul.379733.33785−0.3160%3498−7.8746%3321−12.5362%3341−12.0095%3516−7.4006%
29th Jul.384535.53821−0.6242%3642−5.2796%3214−16.4109%3517−8.5306%3578−6.9441%
30th Jul.414435.84129−0.3620%3859−6.8774%3357−18.9913%3524−14.9614%3874−6.5154%
31st Jul.406136.04054−0.1724%3745−7.7813%3458−14.8486%3642−10.3177%3821−5.9099%