Epileptic seizure detection system and method based on attitude angle signal of portable equipment

The invention discloses an epileptic seizure detection system and method based on attitude angle signals of portable equipment. The method comprises the following steps of: 1, preprocessing pre-marked multi-modal physiological signal data after time calibration; 2, setting different signal combination modes, establishing respective feature matrixes for different signal combinations, firstly performing feature selection operation on each feature matrix, then completing normalization operation to obtain a feature set, inputting the feature set into machine learning for training, and recording the training result of each classifier; 3, performing preprocessing operation on the acceleration signal, the angular velocity signal, the roll angle signal and the pitch angle signal by using a computer, and recording time consumed by different signals in the preprocessing step; and 4, training an LSTM model by using the feature set, and designing an epileptic seizure detection system based on the LSTM through the trained