Electroencephalogram recognition based on feature extraction of wavelet transform and machine learning : Based on Support Vector Machine to identify the validity and accuracy of classification data demonstration

With the development of medical care, predicting disease has become the key to improving patient productivity and quality of life. Therefore, it is particularly important to construct or find an efficient and accurate disease prediction model. In this study, electroencephalogram (EEG) of different disease groups was recorded, and the original EEG signals were preprocessed, analyzed, classified, and feature extracted. The processed EEG data were imported into four network structures for deep learning and testing, and it was verified that Support Vector Machine (SVM) has high accuracy and reliability in several models.

I. Introduction

Introduction: EEG is a very important bioelectrical signal, which reflects a large number of psychological and physiological information of the human body. The recording and analysis of EEG signals are related to brain regions and brain states. The recording, processing and analysis of EEG signals can detect brain activity and extract the required information from it^[3]. It is also one of the important means to understand the information processing process of human brain. The in-depth analysis and processing of EEG signals can provide an important basis for biomedicine, psychology, and clinical disease diagnosis. The research and application of EEG signals mainly focus on the following aspects: brain-computer interface, emotion recognition, diagnosis and prediction of psychological and physiological diseases, and detection and differentiation of brain activity. In the analysis and processing of EEG signals, the processing and analysis of EEG signals is the key technology. In this paper, the above problems are studied, and the original EEG signals are preprocessed, analyzed, classified, feature extracted and other steps, and the processed signals are imported into a variety of network algorithms for memory and recognition. Finally, an algorithm with the highest recognition SVM is compared to prevent diseases^[6].