Adaptive knowledge push method of product intelligent design based on feature transfer

Abstract

To facilitate more effective knowledge utilization in the process of product intelligent design, aiming at the sparsity and cold start issues in knowledge push, an adaptive push method of product intelligent design knowledge based on feature transfer is proposed in the current study. First, based on the research of the current solution strategy, a heterogeneous domain collaborative filtering algorithm model based on feature transfer is constructed, which completed the knowledge transfer from the user behavior domain to the user rating domain, and alleviates the data sparsity issue. Then, for new users, the user-item rating model has been transformed into a tag-keyword relationship model, completing the adaptive reconstruction of the new user rating model, and solves the cold start issue. Finally, the method proposed in this study has been verified by an example, which is proven to be able to accurately push knowledge to new users under the condition of data sparseness while increasing the push rate. It provides a new method for solving knowledge push problems of similar nature.