Multi-dimension semantic dictionary for online intelligence

Abstract

The proposed online intelligence (OI) is to provide popular and well-formatted reports for direct decision support based on online data. Completeness, preciseness, conciseness and smartness are proposed as the basic technical indicators to evaluate an OI application. OI requires similar data processing flow and reporting tools with traditional business intelligence (BI). However, OI collects data basically from online other than provided by enterprise application systems. The online data is usually non-structural or semi-structural. Semantics techniques are required by OI to find and process online data into decisionmaking knowledge automatically in many cases. To this end, a multi-dimension semantic dictionary is proposed to divide online objects' semantics into four semantic aspects, i.e., information semantics, business semantics, application semantics and structural semantics. The four semantic aspects of the semantic dictionary are developed separately and interlinked as a multidimension semantic matrix to conduct matchmaking when find and process online data. In this work, the related concepts and implementation framework of the multi-dimension semantic dictionary are discussed and a real application for illegal drug tracing is provided to illustrate how OI make sense and semantics are developed.