Applying Machine Learning Methods for Prediction of Intellectual Property Patentability

Each of us has heard the word patent, but not everyone thinks about how the institution of intellectual property works, what types of results of intellectual activity it protects and how patenting takes place. At the same time, a large number of entrepreneurs face problems during patenting - it takes quite a long time, does not require small financial costs, and there is always a chance of being refused registration. This paper discusses the possibilities of technical solutions that will facilitate patent applicants to analyze the patentability of their intellectual property in an automated format. This work consists of three parts. The first part analyzes the state and problems of the modern Russian patenting market, as well as identifies barriers hindering the development of patenting. The second chapter focuses on existing technologies and specific systems that solve applied problems in the field of patenting, including simplifying the analysis of the patentability of intellectual property. The third chapter is devoted to the construction of a model that helps to determine the patentability of an application for an invention by analyzing the similarity of the patent description.