psqi_lcga_group_3
CPT_REACTION_TIME_T
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'CPT_REACTION_TIME_T']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.688 0.584 0.605 0.620 0.468 0.507 0.475 0.259 0.620 1.400 22.800
SVM 0.641 0.577 0.596 0.567 0.461 0.459 0.497 0.226 0.567 3.800 16.800
LogisticRegression 0.641 0.558 0.607 0.540 0.405 0.438 0.429 0.216 0.540 6.000 17.400
MLP 0.599 0.566 0.565 0.567 0.463 0.463 0.463 0.185 0.567 1.600 17.200
RandomForest 0.589 0.546 0.531 0.567 0.370 0.363 0.380 0.141 0.567 0.200 19.400
KNN 0.575 0.516 0.548 0.553 0.407 0.672 0.394 0.084 0.553 0.200 22.800
XGBoost 0.567 0.557 0.542 0.573 0.380 0.372 0.389 0.163 0.573 0.200 18.400
DecisionTree 0.490 0.451 0.453 0.453 0.332 0.331 0.338 -0.020 0.453 2.400 18.000
CPT_DPR_T
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'CPT_DPR_T']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.692 0.595 0.613 0.627 0.490 0.547 0.482 0.266 0.627 1.000 22.600
SVM 0.634 0.554 0.576 0.547 0.413 0.422 0.436 0.192 0.547 4.200 17.600
LogisticRegression 0.624 0.533 0.571 0.533 0.410 0.423 0.491 0.205 0.533 6.200 17.800
RandomForest 0.617 0.574 0.562 0.587 0.394 0.387 0.403 0.198 0.587 0.400 17.800
XGBoost 0.600 0.551 0.544 0.560 0.380 0.372 0.388 0.159 0.560 0.400 16.400
KNN 0.595 0.516 0.508 0.533 0.347 0.347 0.353 0.087 0.533 1.000 19.800
MLP 0.585 0.542 0.535 0.553 0.367 0.366 0.370 0.139 0.553 1.200 18.800
DecisionTree 0.523 0.491 0.490 0.493 0.370 0.370 0.370 0.045 0.493 1.600 17.400
CPT_OMISSION_T
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'CPT_OMISSION_T']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
SVM 0.655 0.576 0.609 0.567 0.445 0.457 0.488 0.239 0.567 4.800 17.600
NaiveBayes 0.617 0.585 0.595 0.613 0.477 0.513 0.474 0.241 0.613 1.200 22.000
LogisticRegression 0.616 0.532 0.578 0.527 0.396 0.423 0.451 0.192 0.527 6.200 18.000
KNN 0.579 0.566 0.557 0.600 0.381 0.386 0.394 0.194 0.600 0.200 21.800
RandomForest 0.541 0.516 0.503 0.533 0.347 0.342 0.355 0.084 0.533 0.600 19.200
XGBoost 0.516 0.487 0.493 0.493 0.390 0.484 0.372 0.018 0.493 0.400 17.600
MLP 0.507 0.501 0.496 0.507 0.382 0.388 0.378 0.056 0.507 1.200 18.000
DecisionTree 0.498 0.468 0.470 0.467 0.398 0.404 0.394 0.004 0.467 1.400 16.200
CPT_COMMISSION_T
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'CPT_COMMISSION_T']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.702 0.587 0.604 0.620 0.477 0.520 0.476 0.255 0.620 1.200 22.600
LogisticRegression 0.646 0.523 0.556 0.527 0.407 0.418 0.487 0.182 0.527 5.800 18.200
SVM 0.636 0.533 0.562 0.520 0.414 0.420 0.458 0.166 0.520 5.000 16.600
KNN 0.617 0.559 0.562 0.587 0.376 0.390 0.385 0.182 0.587 1.000 21.600
XGBoost 0.604 0.559 0.546 0.580 0.379 0.375 0.388 0.169 0.580 0.400 19.800
RandomForest 0.595 0.569 0.573 0.587 0.446 0.543 0.430 0.181 0.587 0.400 20.000
MLP 0.582 0.539 0.535 0.547 0.451 0.466 0.442 0.128 0.547 1.200 18.600
DecisionTree 0.513 0.474 0.476 0.473 0.362 0.365 0.360 0.016 0.473 1.400 16.200
CPT_PRS_T
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'CPT_PRS_T']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.662 0.575 0.589 0.607 0.457 0.482 0.467 0.235 0.607 1.600 22.200
SVM 0.638 0.546 0.573 0.540 0.385 0.406 0.391 0.178 0.540 4.400 18.400
LogisticRegression 0.636 0.522 0.569 0.527 0.401 0.424 0.486 0.194 0.527 6.400 18.400
XGBoost 0.599 0.528 0.524 0.533 0.407 0.419 0.401 0.107 0.533 1.000 17.600
RandomForest 0.571 0.514 0.505 0.527 0.349 0.346 0.354 0.079 0.527 0.800 18.800
DecisionTree 0.559 0.540 0.541 0.540 0.435 0.430 0.441 0.142 0.540 2.000 17.200
MLP 0.544 0.504 0.502 0.507 0.382 0.384 0.380 0.068 0.507 1.400 17.200
KNN 0.515 0.538 0.547 0.580 0.360 0.382 0.375 0.145 0.580 0.400 23.400
CPT_HRT_T
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'CPT_HRT_T']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.689 0.587 0.609 0.620 0.472 0.511 0.476 0.259 0.620 1.400 22.600
LogisticRegression 0.642 0.553 0.589 0.547 0.438 0.448 0.508 0.215 0.547 5.400 17.600
SVM 0.623 0.586 0.613 0.567 0.463 0.464 0.502 0.249 0.567 4.400 15.200
KNN 0.620 0.546 0.542 0.580 0.368 0.377 0.381 0.148 0.580 0.200 22.200
RandomForest 0.613 0.561 0.564 0.573 0.443 0.538 0.425 0.160 0.573 0.400 19.200
XGBoost 0.572 0.559 0.554 0.567 0.432 0.455 0.423 0.164 0.567 0.800 18.000
DecisionTree 0.522 0.496 0.494 0.500 0.374 0.375 0.374 0.052 0.500 1.600 18.000
MLP 0.508 0.498 0.486 0.513 0.339 0.334 0.346 0.042 0.513 0.400 19.000
CPT_HRTSD_T
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'CPT_HRTSD_T']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.677 0.587 0.598 0.620 0.439 0.476 0.441 0.251 0.620 1.000 22.400
SVM 0.604 0.532 0.566 0.513 0.396 0.413 0.418 0.158 0.513 5.200 16.600
LogisticRegression 0.596 0.506 0.548 0.500 0.377 0.399 0.434 0.150 0.500 6.400 17.600
RandomForest 0.577 0.537 0.523 0.560 0.364 0.361 0.374 0.121 0.560 0.200 20.200
KNN 0.564 0.538 0.531 0.573 0.361 0.368 0.375 0.133 0.573 0.200 22.200
XGBoost 0.519 0.516 0.502 0.533 0.351 0.345 0.360 0.078 0.533 0.200 19.000
DecisionTree 0.518 0.483 0.481 0.487 0.329 0.329 0.329 0.028 0.487 1.600 17.800
MLP 0.498 0.494 0.485 0.507 0.336 0.333 0.341 0.038 0.507 0.800 19.000
CPT_VAR_T
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'CPT_VAR_T']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.686 0.569 0.576 0.600 0.457 0.482 0.463 0.217 0.600 1.400 22.000
KNN 0.632 0.591 0.582 0.613 0.404 0.406 0.411 0.232 0.613 0.400 20.600
LogisticRegression 0.631 0.517 0.562 0.513 0.397 0.417 0.479 0.178 0.513 6.600 17.600
RandomForest 0.622 0.599 0.584 0.620 0.409 0.404 0.419 0.247 0.620 0.200 19.600
MLP 0.612 0.594 0.591 0.600 0.485 0.492 0.481 0.237 0.600 1.400 18.200
SVM 0.609 0.548 0.591 0.533 0.431 0.447 0.502 0.207 0.533 6.000 16.800
XGBoost 0.570 0.527 0.514 0.540 0.358 0.351 0.366 0.103 0.540 0.400 18.200
DecisionTree 0.513 0.486 0.486 0.487 0.406 0.406 0.406 0.037 0.487 1.600 17.000
CPT_BLOCK_CHANGE_T
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'CPT_BLOCK_CHANGE_T']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.675 0.586 0.602 0.620 0.439 0.480 0.441 0.251 0.620 1.000 22.600
LogisticRegression 0.643 0.593 0.624 0.587 0.496 0.494 0.579 0.278 0.587 4.800 17.200
SVM 0.619 0.557 0.575 0.553 0.449 0.449 0.483 0.190 0.553 3.600 18.000
XGBoost 0.597 0.538 0.524 0.553 0.367 0.358 0.376 0.126 0.553 0.200 18.200
RandomForest 0.592 0.567 0.550 0.587 0.386 0.377 0.397 0.183 0.587 0.000 18.800
MLP 0.584 0.559 0.557 0.567 0.467 0.483 0.457 0.165 0.567 1.200 18.800
DecisionTree 0.538 0.512 0.510 0.513 0.388 0.389 0.386 0.083 0.513 1.400 17.000
KNN 0.515 0.498 0.501 0.540 0.381 0.441 0.382 0.061 0.540 0.600 23.000
CPT_ISI_CHANGE_T
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'CPT_ISI_CHANGE_T']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.649 0.589 0.615 0.627 0.472 0.515 0.479 0.274 0.627 1.400 23.000
LogisticRegression 0.620 0.554 0.593 0.547 0.433 0.446 0.506 0.224 0.547 5.800 17.400
RandomForest 0.586 0.590 0.577 0.607 0.405 0.400 0.413 0.228 0.607 0.400 19.000
SVM 0.577 0.522 0.540 0.513 0.371 0.383 0.377 0.127 0.513 4.000 17.600
DecisionTree 0.561 0.525 0.533 0.520 0.388 0.390 0.392 0.118 0.520 2.600 17.000
XGBoost 0.559 0.559 0.545 0.573 0.384 0.375 0.393 0.167 0.573 0.200 18.000
KNN 0.548 0.539 0.525 0.567 0.363 0.361 0.375 0.130 0.567 0.200 20.800
MLP 0.534 0.520 0.511 0.533 0.353 0.350 0.358 0.092 0.533 0.800 18.800
IGT_NET_TOTAL
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'IGT_NET_TOTAL']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.681 0.580 0.589 0.613 0.438 0.484 0.437 0.232 0.613 0.800 22.400
XGBoost 0.650 0.621 0.618 0.633 0.481 0.528 0.468 0.284 0.633 0.600 19.000
RandomForest 0.645 0.553 0.538 0.573 0.376 0.370 0.386 0.155 0.573 0.200 19.400
LogisticRegression 0.635 0.534 0.588 0.527 0.413 0.440 0.491 0.202 0.527 6.600 17.600
SVM 0.632 0.551 0.578 0.540 0.409 0.422 0.432 0.189 0.540 4.600 17.400
KNN 0.607 0.509 0.496 0.547 0.339 0.340 0.356 0.070 0.547 0.000 22.200
DecisionTree 0.578 0.555 0.551 0.560 0.427 0.439 0.421 0.159 0.560 1.000 17.600
MLP 0.568 0.487 0.487 0.487 0.330 0.329 0.331 0.039 0.487 1.400 16.400
IGT_NET_1
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'IGT_NET_1']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.676 0.589 0.603 0.620 0.494 0.567 0.478 0.247 0.620 0.800 22.400
LogisticRegression 0.642 0.558 0.599 0.553 0.423 0.445 0.473 0.223 0.553 5.400 18.200
SVM 0.609 0.585 0.611 0.580 0.475 0.475 0.535 0.256 0.580 4.400 17.600
XGBoost 0.592 0.540 0.536 0.547 0.418 0.441 0.409 0.127 0.547 0.800 17.800
RandomForest 0.562 0.516 0.501 0.533 0.349 0.342 0.358 0.080 0.533 0.200 18.800
KNN 0.543 0.488 0.472 0.520 0.325 0.322 0.340 0.021 0.520 0.000 21.400
DecisionTree 0.524 0.494 0.494 0.493 0.414 0.420 0.410 0.049 0.493 1.400 16.600
MLP 0.499 0.465 0.457 0.473 0.312 0.308 0.317 -0.014 0.473 0.800 17.800
IGT_NET_2
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'IGT_NET_2']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.671 0.585 0.598 0.620 0.442 0.491 0.441 0.247 0.620 0.800 22.600
LogisticRegression 0.622 0.536 0.586 0.533 0.413 0.438 0.493 0.208 0.533 6.400 18.000
SVM 0.612 0.577 0.593 0.573 0.439 0.443 0.458 0.223 0.573 3.400 18.000
RandomForest 0.610 0.553 0.538 0.573 0.376 0.371 0.386 0.153 0.573 0.200 19.600
XGBoost 0.604 0.577 0.566 0.593 0.395 0.392 0.401 0.203 0.593 0.600 19.200
MLP 0.584 0.519 0.518 0.520 0.387 0.386 0.388 0.099 0.520 1.800 17.400
DecisionTree 0.576 0.543 0.547 0.540 0.440 0.435 0.447 0.151 0.540 2.000 16.200
KNN 0.565 0.518 0.505 0.540 0.349 0.346 0.359 0.086 0.540 0.400 20.000
IGT_NET_3
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'IGT_NET_3']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.680 0.592 0.613 0.627 0.481 0.527 0.480 0.270 0.627 1.200 22.800
SVM 0.661 0.589 0.605 0.587 0.472 0.471 0.505 0.248 0.587 3.400 18.000
RandomForest 0.656 0.596 0.598 0.613 0.462 0.516 0.449 0.236 0.613 0.600 20.400
LogisticRegression 0.649 0.548 0.600 0.533 0.427 0.451 0.500 0.214 0.533 6.400 17.000
XGBoost 0.632 0.568 0.565 0.580 0.442 0.488 0.429 0.179 0.580 0.600 19.000
KNN 0.628 0.551 0.543 0.580 0.374 0.377 0.384 0.153 0.580 0.200 21.400
DecisionTree 0.601 0.569 0.570 0.573 0.419 0.422 0.422 0.194 0.573 2.000 18.600
MLP 0.588 0.530 0.525 0.540 0.405 0.419 0.400 0.109 0.540 1.000 18.800
IGT_NET_4
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'IGT_NET_4']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.696 0.571 0.581 0.607 0.431 0.478 0.431 0.217 0.607 0.800 22.600
SVM 0.652 0.584 0.603 0.580 0.467 0.468 0.501 0.239 0.580 3.600 18.000
LogisticRegression 0.640 0.563 0.614 0.560 0.454 0.474 0.552 0.249 0.560 6.000 18.000
XGBoost 0.625 0.568 0.557 0.580 0.392 0.384 0.400 0.189 0.580 0.400 17.400
KNN 0.619 0.552 0.538 0.573 0.377 0.372 0.386 0.151 0.573 0.200 19.800
RandomForest 0.608 0.534 0.518 0.553 0.364 0.355 0.374 0.114 0.553 0.000 19.000
MLP 0.551 0.464 0.470 0.460 0.383 0.378 0.390 0.010 0.460 2.000 15.600
DecisionTree 0.495 0.458 0.458 0.460 0.309 0.312 0.308 -0.016 0.460 2.000 18.000
IGT_NET_5
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'IGT_NET_5']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.667 0.597 0.623 0.633 0.484 0.535 0.484 0.285 0.633 1.200 23.000
LogisticRegression 0.647 0.561 0.609 0.553 0.440 0.462 0.512 0.233 0.553 5.800 17.800
RandomForest 0.614 0.546 0.529 0.567 0.371 0.363 0.382 0.139 0.567 0.000 19.200
SVM 0.613 0.549 0.571 0.533 0.371 0.391 0.357 0.167 0.533 4.000 17.200
KNN 0.586 0.502 0.486 0.533 0.335 0.331 0.349 0.051 0.533 0.000 21.200
XGBoost 0.579 0.550 0.535 0.567 0.374 0.366 0.384 0.148 0.567 0.200 18.600
MLP 0.558 0.493 0.487 0.500 0.333 0.329 0.337 0.044 0.500 1.000 17.600
DecisionTree 0.557 0.537 0.532 0.547 0.412 0.426 0.406 0.122 0.547 1.000 18.800
IGT_NET_5_MINUS_1
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'IGT_NET_5_MINUS_1']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.690 0.589 0.598 0.620 0.445 0.490 0.442 0.247 0.620 0.800 22.200
LogisticRegression 0.661 0.606 0.668 0.587 0.473 0.506 0.539 0.305 0.587 6.200 17.200
SVM 0.646 0.576 0.591 0.567 0.442 0.443 0.459 0.218 0.567 3.400 17.000
DecisionTree 0.581 0.555 0.553 0.560 0.431 0.454 0.423 0.161 0.560 0.800 17.000
RandomForest 0.580 0.525 0.512 0.540 0.359 0.351 0.368 0.099 0.540 0.200 18.200
KNN 0.568 0.527 0.515 0.553 0.356 0.356 0.367 0.100 0.553 0.200 21.000
XGBoost 0.554 0.513 0.519 0.520 0.409 0.503 0.392 0.070 0.520 0.400 17.600
MLP 0.487 0.487 0.483 0.493 0.374 0.386 0.369 0.029 0.493 1.000 17.800
IGT_DECK_A
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'IGT_DECK_A']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
KNN 0.682 0.566 0.567 0.607 0.381 0.396 0.397 0.199 0.607 0.000 23.000
NaiveBayes 0.680 0.590 0.609 0.627 0.437 0.475 0.443 0.270 0.627 1.200 22.800
MLP 0.671 0.570 0.561 0.580 0.390 0.385 0.395 0.192 0.580 0.800 17.800
SVM 0.657 0.556 0.579 0.547 0.397 0.414 0.400 0.185 0.547 4.000 18.000
RandomForest 0.652 0.599 0.584 0.620 0.409 0.404 0.419 0.247 0.620 0.200 19.600
LogisticRegression 0.630 0.559 0.605 0.547 0.422 0.447 0.471 0.222 0.547 5.800 17.600
XGBoost 0.611 0.551 0.544 0.560 0.376 0.372 0.380 0.155 0.560 1.000 18.000
DecisionTree 0.541 0.499 0.507 0.493 0.337 0.346 0.331 0.069 0.493 2.800 17.400
IGT_DECK_B
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'IGT_DECK_B']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.662 0.576 0.592 0.613 0.426 0.461 0.433 0.240 0.613 1.200 22.800
LogisticRegression 0.636 0.523 0.576 0.520 0.401 0.428 0.483 0.193 0.520 6.800 17.800
SVM 0.616 0.530 0.577 0.507 0.369 0.405 0.370 0.162 0.507 6.200 17.000
RandomForest 0.592 0.527 0.511 0.547 0.358 0.349 0.368 0.101 0.547 0.000 19.000
XGBoost 0.582 0.542 0.532 0.553 0.371 0.364 0.379 0.137 0.553 0.400 17.400
DecisionTree 0.557 0.525 0.519 0.533 0.402 0.414 0.396 0.100 0.533 1.000 18.400
MLP 0.548 0.529 0.525 0.533 0.401 0.407 0.398 0.111 0.533 1.200 17.600
KNN 0.544 0.471 0.455 0.507 0.312 0.308 0.328 -0.013 0.507 0.000 21.800
IGT_DECK_C
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'IGT_DECK_C']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.677 0.571 0.588 0.607 0.453 0.483 0.465 0.234 0.607 1.600 22.600
LogisticRegression 0.647 0.526 0.577 0.527 0.400 0.426 0.486 0.204 0.527 6.800 18.000
SVM 0.639 0.543 0.600 0.520 0.378 0.423 0.377 0.187 0.520 6.400 17.400
XGBoost 0.637 0.586 0.573 0.607 0.399 0.396 0.407 0.223 0.607 0.400 19.800
KNN 0.617 0.551 0.548 0.573 0.372 0.379 0.379 0.159 0.573 1.000 20.800
RandomForest 0.614 0.553 0.538 0.580 0.373 0.370 0.384 0.159 0.580 0.200 20.600
DecisionTree 0.567 0.535 0.533 0.540 0.396 0.396 0.398 0.130 0.540 1.800 18.200
MLP 0.519 0.492 0.485 0.500 0.333 0.329 0.337 0.041 0.500 1.000 17.800
IGT_DECK_D
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'IGT_DECK_D']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.709 0.591 0.606 0.627 0.445 0.498 0.445 0.261 0.627 0.800 22.800
SVM 0.675 0.644 0.662 0.633 0.488 0.493 0.504 0.338 0.633 3.400 17.200
LogisticRegression 0.668 0.576 0.622 0.560 0.450 0.468 0.520 0.255 0.560 6.000 16.800
KNN 0.663 0.571 0.558 0.600 0.388 0.387 0.400 0.194 0.600 0.000 21.000
RandomForest 0.645 0.585 0.571 0.607 0.400 0.395 0.409 0.219 0.607 0.200 19.800
XGBoost 0.635 0.600 0.585 0.620 0.411 0.405 0.421 0.250 0.620 0.200 19.200
MLP 0.628 0.552 0.551 0.553 0.459 0.464 0.455 0.158 0.553 1.400 17.000
DecisionTree 0.606 0.574 0.576 0.573 0.435 0.435 0.436 0.206 0.573 1.600 16.200
WCST_TOTAL_ERRORS_T
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'WCST_TOTAL_ERRORS_T']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.650 0.573 0.582 0.607 0.428 0.463 0.431 0.222 0.607 1.000 22.400
LogisticRegression 0.634 0.575 0.628 0.567 0.448 0.474 0.520 0.262 0.567 6.000 17.800
SVM 0.617 0.558 0.592 0.540 0.415 0.434 0.436 0.199 0.540 5.000 17.000
RandomForest 0.607 0.574 0.557 0.593 0.392 0.383 0.403 0.196 0.593 0.000 18.800
XGBoost 0.601 0.566 0.554 0.580 0.389 0.382 0.397 0.182 0.580 0.400 18.200
DecisionTree 0.561 0.527 0.528 0.527 0.365 0.364 0.365 0.118 0.527 1.600 16.400
KNN 0.555 0.466 0.452 0.500 0.307 0.306 0.323 -0.018 0.500 0.400 21.800
MLP 0.531 0.506 0.505 0.507 0.384 0.386 0.382 0.072 0.507 1.400 16.800
WCST_PCT_ERRORS_T
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'WCST_PCT_ERRORS_T']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.647 0.587 0.604 0.620 0.477 0.520 0.476 0.255 0.620 1.200 22.600
LogisticRegression 0.632 0.558 0.605 0.553 0.437 0.458 0.510 0.232 0.553 5.800 18.000
SVM 0.619 0.597 0.627 0.580 0.449 0.463 0.467 0.260 0.580 4.400 17.000
XGBoost 0.605 0.570 0.555 0.587 0.392 0.381 0.403 0.191 0.587 0.000 17.800
RandomForest 0.602 0.567 0.551 0.587 0.386 0.379 0.395 0.183 0.587 0.200 19.200
MLP 0.533 0.496 0.501 0.493 0.377 0.381 0.376 0.060 0.493 1.400 15.600
KNN 0.518 0.458 0.458 0.487 0.360 0.456 0.351 -0.041 0.487 0.400 21.400
DecisionTree 0.504 0.465 0.471 0.460 0.351 0.350 0.354 0.010 0.460 2.000 15.600
WCST_PERS_RESP_T
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'WCST_PERS_RESP_T']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.668 0.565 0.574 0.600 0.418 0.447 0.425 0.212 0.600 1.200 22.400
LogisticRegression 0.631 0.527 0.562 0.527 0.397 0.413 0.451 0.178 0.527 5.600 18.200
SVM 0.621 0.514 0.541 0.507 0.377 0.392 0.405 0.133 0.507 5.000 17.800
RandomForest 0.572 0.567 0.551 0.587 0.386 0.379 0.395 0.183 0.587 0.200 19.200
XGBoost 0.571 0.531 0.522 0.540 0.363 0.358 0.368 0.115 0.540 0.800 17.800
DecisionTree 0.529 0.487 0.496 0.480 0.338 0.341 0.336 0.055 0.480 2.200 15.200
KNN 0.519 0.429 0.410 0.467 0.277 0.271 0.296 -0.084 0.467 0.600 21.800
MLP 0.515 0.495 0.487 0.507 0.331 0.329 0.335 0.048 0.507 1.200 18.800
WCST_PCT_PERS_RESP_T
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'WCST_PCT_PERS_RESP_T']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.670 0.570 0.583 0.607 0.422 0.454 0.429 0.226 0.607 1.200 22.600
LogisticRegression 0.627 0.515 0.556 0.520 0.397 0.416 0.482 0.177 0.520 6.200 18.400
SVM 0.614 0.526 0.560 0.513 0.388 0.408 0.413 0.150 0.513 5.200 17.600
RandomForest 0.588 0.559 0.542 0.580 0.381 0.372 0.391 0.166 0.580 0.000 19.200
KNN 0.560 0.497 0.494 0.520 0.332 0.339 0.340 0.052 0.520 1.200 21.000
XGBoost 0.558 0.558 0.556 0.567 0.437 0.482 0.425 0.161 0.567 0.600 18.000
MLP 0.524 0.486 0.474 0.500 0.326 0.321 0.333 0.025 0.500 0.600 18.600
DecisionTree 0.508 0.478 0.476 0.480 0.363 0.363 0.363 0.018 0.480 1.600 17.600
WCST_PERS_ERR_T
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'WCST_PERS_ERR_T']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.664 0.565 0.574 0.600 0.418 0.447 0.425 0.212 0.600 1.200 22.400
LogisticRegression 0.630 0.546 0.583 0.547 0.431 0.443 0.505 0.211 0.547 5.400 18.200
XGBoost 0.628 0.566 0.554 0.580 0.387 0.383 0.393 0.180 0.580 0.600 18.800
SVM 0.623 0.549 0.589 0.533 0.400 0.425 0.425 0.196 0.533 5.600 17.400
RandomForest 0.604 0.569 0.553 0.587 0.388 0.380 0.397 0.186 0.587 0.200 18.800
DecisionTree 0.573 0.536 0.541 0.533 0.373 0.373 0.373 0.139 0.533 1.800 15.800
MLP 0.547 0.512 0.505 0.520 0.349 0.344 0.354 0.080 0.520 0.800 17.400
KNN 0.527 0.450 0.431 0.487 0.295 0.289 0.313 -0.056 0.487 0.000 21.800
WCST_PCT_PERS_ERR_T
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'WCST_PCT_PERS_ERR_T']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.676 0.570 0.583 0.607 0.422 0.454 0.429 0.226 0.607 1.200 22.600
LogisticRegression 0.623 0.525 0.563 0.527 0.411 0.425 0.489 0.182 0.527 5.800 18.200
SVM 0.611 0.546 0.576 0.533 0.405 0.421 0.428 0.180 0.533 4.800 17.400
XGBoost 0.588 0.551 0.549 0.560 0.432 0.477 0.419 0.146 0.560 0.600 18.200
RandomForest 0.577 0.566 0.551 0.587 0.386 0.380 0.395 0.182 0.587 0.200 19.400
DecisionTree 0.563 0.529 0.533 0.527 0.402 0.401 0.405 0.127 0.527 1.800 15.800
MLP 0.550 0.513 0.501 0.527 0.349 0.342 0.356 0.077 0.527 0.400 18.200
KNN 0.545 0.472 0.464 0.507 0.310 0.315 0.325 0.009 0.507 1.000 22.000
WCST_NONPERS_ERR_T
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'WCST_NONPERS_ERR_T']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.645 0.583 0.599 0.620 0.435 0.477 0.439 0.251 0.620 1.000 22.800
LogisticRegression 0.635 0.571 0.626 0.560 0.444 0.471 0.516 0.255 0.560 6.200 17.600
SVM 0.617 0.593 0.632 0.567 0.441 0.460 0.459 0.257 0.567 5.200 16.000
RandomForest 0.570 0.541 0.527 0.560 0.369 0.362 0.378 0.130 0.560 0.200 19.200
KNN 0.562 0.522 0.508 0.547 0.352 0.350 0.363 0.090 0.547 0.200 20.600
XGBoost 0.561 0.534 0.523 0.547 0.363 0.358 0.370 0.118 0.547 0.600 18.400
DecisionTree 0.558 0.514 0.522 0.507 0.355 0.360 0.350 0.097 0.507 2.400 16.200
MLP 0.481 0.478 0.486 0.473 0.400 0.402 0.400 0.031 0.473 1.600 15.200
WCST_PCT_NONPERS_ERR_T
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'WCST_PCT_NONPERS_ERR_T']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.672 0.581 0.595 0.613 0.473 0.513 0.472 0.241 0.613 1.200 22.400
LogisticRegression 0.637 0.560 0.611 0.547 0.422 0.452 0.471 0.226 0.547 6.000 17.600
SVM 0.615 0.567 0.605 0.547 0.437 0.452 0.479 0.222 0.547 5.400 16.400
RandomForest 0.567 0.535 0.523 0.553 0.364 0.360 0.372 0.118 0.553 0.400 19.400
KNN 0.558 0.494 0.483 0.527 0.330 0.331 0.344 0.035 0.527 0.200 21.800
XGBoost 0.556 0.514 0.505 0.527 0.349 0.346 0.354 0.079 0.527 0.800 18.800
MLP 0.519 0.517 0.514 0.520 0.395 0.401 0.392 0.090 0.520 1.200 17.200
DecisionTree 0.489 0.451 0.456 0.447 0.308 0.312 0.305 -0.021 0.447 2.200 16.600
WCST_PCT_CONCEPTUAL_T
[資料] 來源=isi_raw_data 目標=psqi_lcga_group_3 列數=150 特徵數=3
[特徵] 使用欄位（前 15）：['BDI_T1', 'BAI_T1', 'WCST_PCT_CONCEPTUAL_T']
[CV] Stratified 5-fold, seed=42 | class_weight=balanced
[聚合] K-fold = weighted | LOSO = weighted
[Leakage check] Class balance
count percent%
psqi_lcga_group_3
0 84 56.0
1 8 5.3
2 58 38.7
[Leakage check] 未發現與目標 |r| ≥ 0.95 的欄位。
=== Basic ML Benchmark (Stratified 5-fold CV) ===
model AUC F1_pos(=1) Prec_pos Rec_pos F1_neg(=0) Prec_neg Rec_neg MCC Accuracy Pred1_mean Pred0_mean
NaiveBayes 0.636 0.568 0.579 0.600 0.458 0.486 0.463 0.216 0.600 1.400 22.200
LogisticRegression 0.623 0.567 0.619 0.553 0.430 0.461 0.476 0.233 0.553 5.800 17.800
SVM 0.614 0.558 0.595 0.540 0.413 0.435 0.434 0.201 0.540 5.200 17.200
XGBoost 0.570 0.559 0.547 0.573 0.383 0.376 0.391 0.168 0.573 0.400 18.200
RandomForest 0.550 0.532 0.516 0.553 0.361 0.354 0.372 0.110 0.553 0.000 19.400
KNN 0.521 0.481 0.467 0.513 0.320 0.318 0.334 0.009 0.513 0.200 21.600
DecisionTree 0.491 0.462 0.458 0.467 0.319 0.315 0.322 -0.011 0.467 1.000 16.800
MLP 0.475 0.431 0.429 0.433 0.294 0.291 0.297 -0.070 0.433 1.000 16.400