Table 2: Comparative analysis of Machine Learning classifier performance: Cross-validation metrics for binary-class and multi-class tasks.

Eighty percent of the electrophysiology dataset was used for cross-validation. The most accurate models for prediction of movement for both multi-class and binary-class are represented with bold text.

	Binary-Class				Multi-Class
Model	Accuracy	Recall	Precision	F1	Accuracy	Recall	Precision	F1
1. Ada Boost Classifier	81.36	88.12	87.36	87.24	60.45	60.18	64.6	60.32
2. CatBoost Classifier	85.91	94.38	87.81	90.73	75	75.7	77.49	74.53
3. Decision Tree Classifier	78.64	82.65	88.1	84.86	58.64	59.1	61.34	58.6
4. Dummy Classifier	73.18	100	73.18	84.51	38.18	33.33	12.73	18.41
5. Extra Trees Classifier	83.18	92.54	86.12	88.98	75.45	75.58	77.32	74.47
6. Extreme Gradient Boosting	82.27	89.38	87.74	88.08	69.09	69.58	70.97	67.83
7. Gradient Boosting Classifier	88.64	93.12	92.12	92.34	72.73	72.87	74.45	72.23
8. K Neighbors Classifier	78.18	88.86	83	85.62	56.82	57.8	60.22	56.12
9. Light Gradient Boosting Machine	86.36	91.91	90.33	90.78	70	70.05	71.01	69.11
10. Linear Discriminant Analysis	80.45	86.36	87.32	86.53	66.82	67.83	69.17	66.1
11. Logistic Regression	80.91	87.65	86.58	86.91	61.36	61.87	62.74	60.76
12. Naive Bayes	78.18	78.2	91.25	83.69	43.64	42.51	47.83	40.94
13. Quadratic Discriminant Analysis	80.45	97.54	80.23	87.98	67.73	65.53	74.27	65.51
14. Random Forest Classifier	83.64	93.75	85.82	89.29	72.73	72.38	74.3	71.72
15. Ridge Classifier	80.45	86.99	86.74	86.59	63.64	64.15	66.32	62.58
16. SVM - Linear Kernel	71.36	79.78	84.03	78.74	50.45	51.16	51.03	47.34