Epilepsy Detection Using Random Forest Classification Based on Locally Linear Embedding Algorithm

Epilepsy is a common disease of the brain nervous system. The key to epilepsy surgery is to locate the epileptic foci. Research shows that they can be detected by magnetoencephalographic (MEG) data. The Random Forest Classification model based on Locally Linear Embedding (LLE) is used to clean the obtained brain nerve data, and then the dimensionality reduction operation is performed through the LLE algorithm. Finally, the optimized Random Forest algorithm is used to construct a classification model that can obtain a higher average accuracy rate. This experiment shows that the average accuracy of the optimized model is improved to 0.95, which is more effective than the decision tree algorithm, support vector machine algorithm, k-neighbor algorithm and Random Forest algorithm without any optimization. Therefore, it can classify the epilepsy lesion more accurately with a strong generalization ability and robustness.