Systems for Economics Monitoring

 

The English release of the Semantix linguistic processor is being tuned to the Financial knowledge extraction and analysis and semantic Web search for the relevant information in Economics sphere. This release is intended for intelligent support of business solutions. The primary investigation stage requires cognitive linguistic analysis of the subject area with the subsequent design of the domain-oriented ontology and its incorporation into the general ontology of Semantix. The development of the domain-specific vocabularies can be partially supported by the learning facilities of the linguistic processor.

The system performs semantic objects extraction from natural language texts and automatic formation of semantic networks. The networks are populated by objects of different types and their interconnections: banks, companies, addresses, stock, shares, values, stock exchanges, etc. The objects are extracted from documents with their features and relations. On the basis of each document a special form of semantic network is built reflecting its semantic structure. Such networks are mapped into XML- files which constitute knowledge bases. Semantic search results in semantic presentations for deciding logical analytical problems and for automatic filling of relational databases.

     Three modes of semantic presentations are supported by the system: 1) List of objects types with references to particular content items, e.g. Type: Mentioned Organizations; Content: Bank of China; Source Text: “Bank of China”; 2) Graph presentation; 3) XML presentation of text.

 

Graph of extracted objects and their links and interconnections: