Table 4 Nonparametric test for classification metrics of 7 ML models (Kruskal–Wallis test, M (QL, QU))
ET | Catboost | RF | LightGBM | GBC | XGBoost | DT | |
|---|---|---|---|---|---|---|---|
Accuracy | 0.8571(0.8285,0.8605) | 0.8571(0.8285,0.8605) | 0.8372(0.8333,0.8605) | 0.8140(0.8048,0.8393) | 0.8237(0.7849,0.8343) | 0.8237(0.7811,0.8343) | 0.7093(0.6154,0.7486)*#â–³ |
AUC | 0.9110(0.8950,0.9315) | 0.9010(0.8936,0.9278) | 0.9039(0.8955,0.9289) | 0.9022(0.8915,0.9273) | 0.9063(0.8938,0.9187) | 0.9085(0.8908,0.9305) | 0.7821(0.7070,0.8149)*#△□◇※ |
Recall | 0.8428(0.8207,0.8550) | 0.8428(0.8117,0.8565) | 0.8338(0.8281,0.8468) | 0.8039(0.7935,0.8274) | 0.8120(0.7836,0.8301) | 0.8079(0.7742,0.8292) | 0.7033(0.5938,0.7434)*#â–³ |
Prec | 0.8618(0.8383,0.8780) | 0.8606(0.8312,0.8742) | 0.8397(0.8349,0.8721) | 0.8212(0.8104,0.8402) | 0.8321(0.7899,0.8422) | 0.8307(0.7944,0.8422) | 0.7095(0.6361,0.7518)*#â–³ |
F1 | 0.8559(0.8294,0.8602) | 0.8539(0.8281,0.8596) | 0.8354(0.8319,0.8603) | 0.8121(0.8039,0.8337) | 0.8223(0.7855,0.8344) | 0.8202(0.7822,0.8348) | 0.7073(0.6142,0.7484)*#â–³ |
Kappa | 0.7844(0.7482,0.7905) | 0.7843(0.7422,0.7900) | 0.7543(0.7489,0.7903) | 0.7194(0.7048,0.7576) | 0.7346(0.6762,0.7510) | 0.7348(0.6705,0.7505) | 0.5614(0.4170,0.6221)*#â–³ |
MCC | 0.7910(0.7466,0.7992) | 0.7870(0.7438,0.7947) | 0.7568(0.7505,0.7956) | 0.7239(0.7065,0.7631) | 0.7391(0.6777,0.7571) | 0.7392(0.6737,0.7571) | 0.5630(0.4237,0.6242)*#â–³ |