The invention relates to the technical field of
machine learning,
deep learning,
data mining and the like, in particular to a hierarchical Attention
deep learning model course recommendation method. According to the method, dynamic changes of user interests are modeled by utilizing user sequential behavior data and using LSTM, and long-term and short-term preferences of a user are obtained by constructing a hierarchical Attention structure, so that advanced mixed representation of the user is generated, and user
personalization and accuracy of a recommendation result are improved. The method specifically comprises the following steps: performing screening and preprocessing by using original
online learning related behavior data, dividing the sequential behavior of a user into sessions, and then
processing three types of data of fine
granularity (information user ID and course ID) and coarse
granularity information (course type) by using an embedding layer and a full connection layer to obtain user vector representation; capturing interaction and evolution of different historical session interests of a user by applying LSTM (
Long Short Term Memory) to obtain a serialized interest vector, and inputting the interest vector into an
Attention network to obtain long-term interest representation of the user; inputting the recent behavior data and the long-term interest representation of the user into a second-layer
Attention network to obtain a mixed interest representation of the user; and finally, carrying out inner product on the mixed interest representation and the course vector representation of the user, taking an obtained value as a
score of the candidate item, and sorting the scores of the candidate item to obtain a recommendation
list for carrying out personalized recommendation on the student.