Patent collaborative filtering recommendation approach based on patent similarity
Xiang Ji, GU Xin-jian, Feng Dai, Jixi Chen, Chengyi Le · 2011
In order to help enterprises deal with the situation of sharp increase in patent amount and improve enterprises' capabilities for patent utilization, the information filtering technology was introduced into patent management field. A patent collaborative filtering recommendation approach based on patents similarity and its corresponding algorithm was presented. This approach solves the problem of user rating data sparsity successfully: firstly, it calculates similarity between patents based on the concept of patent model tree, then predicts patent score with patents similarity, fills user rating matrix, and finally it calculates users' similarity to finish patent recommendation. Preliminary experimental results showed that the patent collaborative filtering recommendation approach based on patent similarity has a good performance in recommending patents and patent management system employing this approach can improve enterprises' capabilities in patent utilization.