While managing the innovation activity it is necessary to base on the reliable sources of scientific and technical information, one of which is the patent research. However, the existing variety and scale of the patent databases lead to the necessity of the elaboration of the instrument that allows processing the large volumes of the patent information in limited terms. In these conditions, it is necessary to use the machine learning technology that allows creating the information base of the management decisions. In the present paper the algorithm for processing patent data is suggested the purpose of which is the quality increase of the conducted patent research. The essence of the algorithm is that all necessary patents are ranking according to the relevance criterion and the researcher analyses already necessary patents. The suggested instrument allows decreasing time for processing in 4.5 times. The present paper envisages the practical realization of the algorithm on the example of a gravity driven power generator.

JEL Codes:

Keywords: Innovation activity; patent analytics; machine learning technology; a gravity driven power generator.