A robust epilepsy diagnosis method based on SDAE and CatBoost

With the development of deep learning, it is found that it can automatically learn rules from complex data sets and find features that are difficult to find manually. The robust epilepsy diagnosis method based on sdae and catboost is a technology that can be used to diagnose the types of epilepsy. This method includes two steps: first, we use sdae (seizure diagnosis algorithm) to classify all seizures in the database into different seizure types; Then, we use the catboost algorithm to identify new biomarkers for each seizure type. In this paper, we combine the two methods and propose an effective robust seizure classification scheme. We prove that if only catboost algorithm is used without any other technology, such as sdae or K-means clust.