In vivo electrophysiology recordings and computational modeling can predict octopus arm movement

Table 4 Comparison of models across time frames

Best Model	Testing Accuracy	Accuracy (Cross-validation)	Dataset	Target
Extra Trees Classifier	85.45	86.82	Spatial–temporal features dataset with 50 ms bin size	Binary Class
Extra Trees Classifier	74.55	74.55	Spatial–temporal features dataset with 50 ms bin size	Multi-Class
Gradient Boosting Classifier	83.64	88.64	Spatial–temporal features dataset with 100 ms bin size	Binary Class
Extra Trees Classifier	72.73	75.45	Spatial–temporal features dataset with 100 ms bin size	Multi-Class

The outperforming models were tested on electrophysiology signals binned into 100 ms and 50 ms features. Comparing the accuracy of the results from the cross-validation shows that 100 ms features have led to 1.82% higher accuracy in Binary-class and 0.9% higher accuracy in the Multi-class compared to the 50 ms features

ISSN: 2332-8886