Research on Knowledge Recommendation Based on Weighted Directional Tripartite Graphic Network Structure
Juanli Zheng, Fuyi Wei · 2017
Using the material diffusion algorithm and the heat propagation algorithm, introduced the concept of directional complete bipartite graph and directional star graph. User similarity and tag similarity are transformed into the weighted problem of directional complete bipartite graph, combined with the network active period and the label time decay characteristic is given weighting formula. Using the multiplication principle and additive principle in statistics, by using weighted directional tripartite graph, with a series of comprehensive empowerment formulas, modified similarity and scoring formulas were obtained. The scheme of knowledge recommendation algorithm is obtained. This method theoretically improves the accuracy of the knowledge recommendation model and can be applied to push tags to target users or push labels to the target project.