Construction of dynamic update and adaptive prediction model for user profile based on time series analysis

The rapid development of Internet technology has made the phenomenon of “information overload” more and more obvious, and it has become more and more difficult for users to filter out useful information from the huge amount of information. The deep forest model is used in the study to predict the establishment of labels in the user profile system. Furthermore, the model utilizes a time-attention system to update the user profile dynamically and develops an adaptive weight combination strategy to enhance the prediction accuracy of the combination prediction model. According to the model performance analysis, three models, RF, ET, and XGB, were selected to form the cascade forest module of deep forests. The prediction accuracy of this paper’s method for the labels in the user portrait system is 92.3%, and the prediction performance is good. After being applied to students’ personalized learning path recommendation system, most students recognize the effect of the recommendations.