Evaluating Machine Learning Algorithms on BCI Dataset

Brain Computer Interface (BCI) or Brain Machine Interface (BMI) is evolving rapidly, as machine learning algorithms are getting improved. Every algorithm uses its own approach to train and test on particular dataset. FingerFlex is one of the standard EEG dataset related to finger movement of nine subjects. Multiple algorithms are implemented on this dataset to evaluate best performance algorithm such as Linear Regression (LR), decision Tree (DT), Bayesian Ridge (BR), Extra Tree Regression (ETR), Light Gradient Boosting (LGB), Extreme Gradient Boosting (XGB) and K Neighbor Regression (KNR). The final result is presented with highest achieved correlation value as 0.99 by using the KNR and 0.19 by using the linear regression algorithm.