A Novel Graph Decomposition Approach to the Automatic Processing of Poorly Formalized Data: Innovative Ideas – A Management Case Study

In the following paper we present a novel approach to unstructured data processing by imposing a hierarchical graph-based structure on the data and decomposing it into separate subgraphs according to optimization criteria. In the scope of the paper we also consider the problem of automatic classification of textual data for the synthesizing the hierarchical data structure. The proposed approach uses textual information on the first stage to classify ideas, innovations, and objects of intellectual property (OIPs) to construct a multilayered graph. Numerical criteria are used to decompose constructed graph into separate subgraphs. In the scope of the research we apply the developed approach to the innovative ideas in a management case study. The research has been conducted in the scope of a joint research project with financial aid of Ministry of Education and Science of Russian Federation RFMEFI57314X0007.