A Hybrid Knowledge Push Method Based on Trust-Aware and Item-Cluster Oriented to Product Design

Abstract

A knowledge push system as a proactive method of knowledge management, and is an effective way to push knowledge to designers in the product design process. Knowledge push can help to improve design efficiency and quality, but the knowledge pushed to designers is less applicable in engineering scenarios. To improve the quality of knowledge push, we propose the trust-aware and item-cluster strategies to extract the two nearest neighbors of the designers and the design knowledge. Next, we adapt a fusion model for the two nearest neighbors, which is solved based on the gradient descent algorithm. Finally, the rating prediction that the designer gives to the knowledge component is the decision point of the knowledge push system, where we propose our final hybrid knowledge push method. A knowledge push system of CNC machine tools design is used to confirm that our method outperforms the other widely adopted methods.