Table 7 Training and testing data size and method used to split training and testing data
| Â | Training and testing datasets | ||
|---|---|---|---|
Training data size | Testing data size | Method | |
Geng et al. (2014)* [26] | 100 200 | 1000 1000 | Independent samples from DGM |
Golmakani et al. (2020) [31] | 450 720 | 50 80 | 10-fold cross-validation |
Gong et al. (2018) [27] | 200 400 500 600 800 1000 | 200 400 500 600 800 1000 | Independent samples from DGM |
Hu and Steingrimsson (2018) [28] | 200 500 | 1000 1000 | Independent samples from DGM |
Katzman et al. (2018)* [29] | 4000 | 1000 | Independent samples from DGM |
Lowsky et al. (2012)* [25] | 500 1000 3000 7500 | 13525 13525 13525 13525 | Independent samples from DGM |
Omurlu et al. (2009) [24] | 50 100 250 500 | 50 100 250 500 | Unclear |
Steingrimsson and Morrison (2020) [32] | 250 500 1000 1500 3000 | 250 500 1000 1500 3000 | Independent samples from DGM |
Wang and Li (2019) [30] | 150 | 150 | Two-fold cross-validation |
Xiang et al. (2000) [23] | 100 200 | 100 200 | Randomly split whole sample into equal training and testing sets |