Development of the patent array analysis module based on the “Problem-Solution” model

Abstract

Abstract In the modern world, there is a description of many technical solutions in different types. A patent is one of these types. Analysis of the patent array based on the “Problem-Solution” model allows not only to analyze existing technical solutions, but also to improve the characteristics of individual elements of a technical object and, accordingly, the technical object (TO) itself. To do this, you need to detail the structural-functional structure of the TO, i.e. define the elements (components) of TO and the functions of the TO elements. To extract descriptions of functions of a technical object, we use the semantic construction “Subject-Action-Object” (SAO). If “Action” is an attributive verb, such as “reduce”, “improve”, “increase”, etc., then “Subject” can affect “Object” as a problem or need. The SAO structure can be organized in the “problem-solution” format, where the “Object-action” (AO) forms the problem or need, and the “Subject” (S) sets the solution. Structured data will be extracted from natural language text using the Yandex Tomita-parser technology. Then, after all the procedures, it is advisable to fill the database with the information obtained for convenient analysis and search.