Summary of 5_LightGBM

<< Go back

LightGBM

• n_jobs: 6
• objective: multiclass
• num_leaves: 15
• learning_rate: 0.05
• feature_fraction: 0.8
• bagging_fraction: 0.5
• min_data_in_leaf: 50
• metric: custom
• custom_eval_metric_name: f1
• num_class: 6
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True
• random_seed: 42

Optimized metric

f1

Training time

27.5 seconds

Metric details

	0	1	2	3	4	5	accuracy	macro avg	weighted avg	logloss
precision	0.970812	0.96534	0.971355	0.98567	0.992358	0.997917	0.980461	0.980575	0.98054	0.066517
recall	0.98619	0.962868	0.968878	0.990741	0.993986	0.976057	0.980461	0.979787	0.980461	0.066517
f1-score	0.97844	0.964103	0.970115	0.988199	0.993171	0.986866	0.980461	0.980149	0.980466	0.066517
support	2462	1562	1960	1944	1829	1963	0.980461	11720	11720	0.066517

Confusion matrix

	Predicted as 0	Predicted as 1	Predicted as 2	Predicted as 3	Predicted as 4	Predicted as 5
Labeled as 0	2428	8	4	19	0	3
Labeled as 1	2	1504	50	2	4	0
Labeled as 2	7	46	1899	2	5	1
Labeled as 3	16	0	1	1926	1	0
Labeled as 4	5	0	1	5	1818	0
Labeled as 5	43	0	0	0	4	1916

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Precision Recall Curve

<< Go back