Table 4:

10-fold cross-validation performance metrics for different window sizes with an embedding dimension of 21. The highest values in each category are highlighted in boldface. MCC: Matthew’s Correlation Coefficient; SN: sensitivity, SP: Specificity; ACC: Accuracy.

Size	CNN				LSTM
	MCC	SN	SP	AC C	MCC	SN	SP	AC C
51	0.52	0.71	0.81	0.76	0.44	0.66	0.77	0.72
45	0.52	0.72	0.80	0.76	0.44	0.66	0.77	0.72
39	0.53	0.73	0.80	0.76	0.45	0.66	0.78	0.72
33	0.52	0.71	0.81	0.76	0.46	0.66	0.79	0.73
27	0.52	0.70	0.81	0.76	0.46	0.67	0.78	0.73
21	0.50	0.71	0.79	0.75	0.46	0.65	0.80	0.73
15	0.49	0.70	0.79	0.74	0.44	0.66	0.77	0.72
9	0.46	0.69	0.77	0.73	0.42	0.65	0.76	0.71