Detection of Epileptic Seizures from Logistic Model Trees
V. Nageshwar, P. Venkateswara Rao, C. Sarika, K. Manusha, Y. P. Deepthi · Atlantis Highlights in Computer Sciences/Atlantis highlights in computer sciences · 2023
Electroencephalogram (EEG) signals from patients can be analyzed to identify the neurological condition known as epileptic seizures.Because to the large dimensionality of the data and the existence of noise and artifacts, detecting epileptic seizures from EEG signals is a complicated process.In this article, we present a novel method for epileptic seizure detection using principal component analysis (PCA) as the feature extraction method and logistic model trees (LMT) as the classification method.By converting the original features into a lowerdimensional space, PCA is a frequently used feature extraction approach that lowers the dimensionality of the data.LMT is a decision tree-based machine learning method with the advantage of being able to incorporate linear models into its decision tree structure.It can handle non-linear connections between variables.We used a publicly accessible dataset of EEG signals captured from epilepsy patients for our study.The most crucial elements from the EEG signals were then extracted using PCA.Then, using the extracted features to train an LMT classifier, we assessed the classifier's performance using a variety of measures, including accuracy, precision, recall and f1_score.The proposed method has an accuracy of 95.33%, a precision of 93%, f1_score 92.5% and recall of 92% according to our experimental findings.These findings show how successfully the suggested method can identify epileptic seizures from EEG signals.The suggested method may be helpful for creating a real-time seizure detection system that will help in epilepsy diagnosis and care.