Research Article
Improved Multiclassification of Schizophrenia Based on Xgboost and Information Fusion for Small Datasets
Table 1
Classification metrics of different methods.
| | Methods | Accuracy | Precision | Recall | F1-scores | AUC |
| | LR (GMV) | 64.8611% | 65.0509% | 64.8611% | 0.6237 | 0.8160 | | LR (ALFF) | 65.9722% | 65.5394% | 65.9722% | 0.6388 | 0.7743 | | SVC (GMV) | 61.5278% | 51.5602% | 61.5277% | 0.5508 | 0.7403 | | SVC (ALFF) | 67.0833% | 59.4147% | 67.0833% | 0.6114 | 0.6114 | | KNN (GMV) | 63.4722% | 61.4259% | 63.4722% | 0.5899 | 0.7174 | | KNN (ALFF) | 58.6111% | 67.3175% | 58.6111% | 0.5601 | 0.6731 | | NN (GMV) | 55.4167% | 56.4749% | 47.2222% | 0.5425 | 0.8066 | | NN (ALFF) | 64.4444% | 67.0764% | 60.8333% | 0.6266 | 0.7618 | | NB (GMV) | 53.8889% | 55.8657% | 53.8889% | 0.5216 | 0.7389 | | NB (ALFF) | 68.8889% | 74.0139% | 68.8889% | 0.6781 | 0.8308 | | CART (GMV) | 45.8333% | 47.7870% | 51.9444% | 0.4721 | 0.6375 | | CART (ALFF) | 60% | 57.6075% | 56.6667% | 0.5667 | 0.6542 | | RF (GMV) | 62.6389% | 60.0883% | 62.6389% | 0.5874 | 0.7722 | | RF (ALFF) | 65.6944% | 57.1042% | 65.6944% | 0.5943 | 0.7969 | | Xgboost (GMV) | 62.6389% | 60.4445% | 62.6389% | 0.5888 | 0.7958 | | Xgboost (ALFF) | 68.0556% | 62.9610% | 68.0556% | 0.6364 | 0.8351 | | DNN | 62.7083% | 40.3380% | 62.7083% | 0.5152 | 0.6723 | | Our classifier (fusion) | 73.8889% | 65.4242% | 73.8889% | 0.6746 | 0.8524 |
|
|
