Patent personalised recommendation method based on fusing co-occurrence network and point mutual information

In the emerging high-tech industry, the number of patents is growing particularly rapidly. In this background, timely and accurate identification of patents that are closely related to enterprises and have significant influence for realising patent transformation and promoting enterprise development. In this paper, we propose a recommendation method based on co-occurrence network and Point Mutual Information coefficient (PMI). Through experiments on the patent texts in the communication industry, this paper finds that the patents recommended are highly compatible with the development direction of the enterprise, which can provide high value for the development of the enterprise. It verifies that the method of this paper is valid in the field of patent recommendation, and provides new ideas for improving the conversion rate of patents and promoting the application of scientific and technological achievements.