Analyzing the difference between deep learning and machine learning features of EEG signal using clustering techniques

The performance of many important Brain Computer Interface (BCI) applications depends on its classification system. In former decades of BCI, machine learning classification algorithms serve this purpose. But in the recent picture of BCE, deep learning replaces the position of conventional machine learning algorithms. The research papers of BCI, where deep learning is used for the classification, claim to get better accuracy for the system compares to the use of machine learning. Though the machine learning classification algorithms are specifically designed for the purpose of classification, deep learning achieves better results by using a single layer of Softmax for the job of classification. The key behind this achievement of deep learning is its extracted features from several non-linear hidden layers. These features are far better for the task of classification than the handcrafted features used in the machine learning algorithms. In this work, we have analyzed the difference between these two different kinds of features with the support of clustering techniques.