Clustering Analysis and Visualization of TCM Patents Based on Deep Learning

In the process of medicine innovation, pharmaceutical enterprises tend to seize the intellectual property highland actively. They engage in research and development independently, apply for patents for core technologies, or take the initiative to acquire patents from others. Before applying for patents by their own efforts or purchasing patents from others, pharmaceutical companies need to search for related patents in the patent pool and make a comparative analysis of them, in order to find technology blank areas as R&D objectives, or find valuable patents as potential acquisition targets. In this paper, we use deep learning technology and propose a semantic-based clustering algorithm for Traditional Chinese Medicine (TCM) patents, discarding the traditional literal-based text clustering method. We also give a visualization method for TCM patents, so as to facilitate pharmaceutical enterprises to intuitively understand the relevant patents.