Abstract: D-matrix is a standardized diagnostic model which is used to catch the fault system data and its causal relationship at the hierarchical system-level. Construction of D-Matrix by using the data source views the overall data by saving the entire database. Further it parses that data scanning which requires more memory and makes the process time consuming. Is describes construction and updation of D-Matrix by auto mining the unstructured repair verbatim ( written in unstructured text) data collected during fault diagnosis using document pre-processing, term extractor and phrase merging. The system composes the fault diagnosis ontology comprising of dependencies and relationships currently saw in the fault diagnosis domain of interest and then text mining algorithms make utilizing of ontology concept to identify the necessary artifacts, like as failure modes, parts, symptoms and conditions from the unstructured repair verbatim content. The proposed scheme is equipment as a prototype tool and accepted by utilizing real time information gathered from the automobile domain of interest.

Keywords: Data Mining, fault analysis, fault diagnosis, information retrieval, text processing