The Formation of Metrics of Innovative Potential and Prospects for the Task of Technological Forecasting

Abstract

n today's rapidly developing technological world, new ideas, inventions and developments appear daily. At the same time, individual technologies may have common features, through the use of similar methods, modification and expansion of existing technologies, or solving a common problem for which developments are being created. Patenting is often used to preserve these ideas and protect intellectual property. During the development of the program for the analysis of the patent array to obtain criteria assessments of innovation potential and prospects, enshrined in patent high—tech technical systems and technologies, the subject area — patent array was investigated, methods for analyzing texts in natural language and various options for determining the criteria of innovation potential were considered. As a result, algorithms were developed to obtain the following criteria for assessing the innovative potential of a patent: the mass nature of the subject of this technology for the current year and the estimated frequency of occurrence for the next, the economic characteristics of the patent holder's company, the potential citation of the patent. These criteria are determined based on the analysis of texts and patent data by means of clustering, classification, regression analysis, normalization of the name of the patent holder. The developed algorithms have been tested on patents issued by the US Patent and Trademark Office, as well as on Google Patents.