Research Article

A Short-Term Wind Speed Forecasting Hybrid Model Based on Empirical Mode Decomposition and Multiple Kernel Learning

Table 5

Multistep forecasting results with confidence intervals.

EMD-SVMEMD-MKL
1-step prediction2-step prediction3-step prediction1-step prediction2-step prediction3-step prediction

−2.8694−2.8616−3.5215−1.2349−1.2083−1.3172
−2.9508−4.0046−4.4386−1.6884−2.0611−2.0446
−1.8067−2.8289−3.6129−1.4967−1.9216−2.4128
−2.2733−3.0428−3.8066−1.7741−2.0374−2.1500
−2.7841−2.5692−3.1499−1.4236−1.6013−1.8308
−2.8305−2.9841−3.4719−1.3805−1.4735−1.4825
−1.3578−1.8935−2.3102−1.0058−0.9542−1.2542

0.40180.84930.91810.37810.60350.6765
0.72191.21861.92460.60730.73310.9268
2.65432.70813.09320.97861.09131.1415
3.11042.98393.42461.32371.27191.4514
1.99042.80543.35261.32161.37111.6346
2.80942.82933.34771.40831.51311.7677
2.88963.30324.26791.39661.95442.1963