Semantic processing of unstructured textual data based on the linguistic processor PullEnti

The paper presents the method for creation of knowledge extraction systems based on the approach employing the software tool system PullEnti comprising the algorithms for morphological and semantic-syntactical analysis which makes it possible to extract entities of certain types from natural language texts (persons, organizations, locations, and other target semantic objects). The PullEnti system uses dynamically connected components (plugins) which makes it possible to activate various functions without recompiling. This is how the semantic analysis unit is incorporated. During the analysis, the semantic units (tokens) are established, which are typed phrases: text, numerical data, etc. Examples of implemented projects for different subject areas are given.