Efficient Detection of Focal EEG Signals Using Shearlet Transform and Random Forest Classifier

Focused and non-focal signals are the two categories that are used to categorize the electroencephalogram (EEG) signals that are produced by an ERP sensor. In individuals, epilepsy is caused by an aberrant signal that originates from a specific area of the brain. This incorrect signal is referred to as a focused signal. It is also referred to as focal seizure in other languages. Generally speaking, the signal that originates from the brain area is considered to be a non-focal signal. An unexpected pattern may be seen in both the focal & non-focal EEG signals. For this reason, it is essential that we create methods that can detect epilepsy syndrome at an earlier stage. Utilizing machine learning and methods of deep learning, this study effort proposes a unique approach for the identification of epilepsy. The purpose of this study's operate is to suggest three distinct techniques for automated categorization of electroencephalogram (EEG) signals through focal or non-focal signal categories. These approaches are as follows: Shearlet transform based Random Forest (STRF) classification approach. Compared to RF using DWT Transform and Contourlet Transformation, which is the suggested approach, RF using Shearlet Transform, provides a grossly superior 99.2% accuracy. Emphasizing its resilience in identifying both real positives and negatives, relative to the other approaches. Because of this, it is the method that outperforms the others that were considered.