Research Article

Online Sequential Prediction for Nonstationary Time Series with New Weight-Setting Strategy Using Extreme Learning Machine

Table 5

Comparative results of three models.

ā€‰ ELM OS-ELM LW-OSELM

Training time(s) 0.0125 0.0998 3.3961
Test time(s) 0.0031 0.0094 0.0012
Training error 0.1706 0.1721 0.1705
Test error 0.1811 0.1808 0.1739