A method for matching optimal parameters of cognitive training based on an LSTM model

Abstract

This invention relates to the intersection of cognitive training intervention and artificial intelligence, specifically to a method for matching optimal parameters for cognitive training based on an LSTM model. The method includes: designing features and methods for collecting user cognitive training log data; automatically evaluating user cognitive abilities based on a cognitive unit scoring method; and constructing a dataset based on the above design. The dataset is preprocessed to reconstruct a sequence dataset of cognitive training interventions as training samples. These training samples are then input into a depth-optimal difficulty prediction model designed based on an LSTM framework for training. Finally, the trained prediction model is used to predict the difficulty coefficient of cognitive training and to evaluate model performance. This invention, by training an LSTM prediction model to determine the difficulty of the next training session based on historical data from user cognitive training, ensures that users consistently undergo cognitive training at a challenging level of difficulty and intensity.

Images