Table 3 Accuracy, rain bias, and snow bias for the three machine learning phase partitioning methods (PPMs): the artificial neural network (ANN), random forest (RF), and XGBoost (XG) applied to the two datasets
Performance Metric for Machine Learning Models | Absolute Difference of Performance Metric to Benchmark Values | |||||||
|---|---|---|---|---|---|---|---|---|
Dataset | PPM | Benchmark Comparison | Accuracy (%) | Snow Bias (%) | Rain Bias (%) | Accuracy (%) | Snow Bias (%) | Rain Bias (%) |
Crowdsourced | ANN | Best | 89.2 | 3.8 | −8.6 | 0.5 | −0.7 | −1.6 |
RF | Best | 88.3 | 3.5 | −7.9 | −0.4 | −1 | −2.3 | |
XG | Best | 88.8 | 4.7 | −10.6 | 0.1 | 0.2 | 0.4 | |
ANN | Average | 89.2 | 3.8 | −8.6 | 3.2 | −1.7 | −3.9 | |
RF | Average | 88.3 | 3.5 | −7.9 | 2.3 | −2 | −4.5 | |
XG | Average | 88.8 | 4.7 | −10.6 | 2.8 | −0.8 | −1.8 | |
Synoptic | ANN | Best | 92.8 | 1.5 | −1.4 | -0.3 | −4.7 | −4.1 |
RF | Best | 93.7 | 3.6 | −3.2 | 0.6 | -2.6 | −2.3 | |
XG | Best | 93.3 | 5.4 | −4.8 | 0.1 | -0.8 | −0.7 | |
ANN | Average | 92.8 | 1.5 | −1.4 | 0.8 | −5.9 | −5.2 | |
RF | Average | 93.7 | 3.6 | −3.2 | 1.7 | −3.8 | −3.3 | |
XG | Average | 93.3 | 5.4 | −4.8 | 1.3 | −2 | −1.7 | |
