Abstract
Analysis of warranty based big data has gained considerable attention due to its potential for improving the quality of products whilst minimizing warranty costs. Similarly, customer feedback information and warranty claims, which are commonly stored in warranty databases might be analyzed to improve quality and reliability and reduce costs in areas, including product development processes, advanced product design, and manufacturing. However, three challenges exist, firstly to accurately identify manufacturing faults from these multiple sources of heterogeneous textual data. Secondly, accurately mapping the identified manufacturing faults with the appropriate design information and thirdly, using these mappings to simultaneously optimize costs, design parameters and tolerances. This paper proposes a Decision Support System (DSS) based on novel integrated stepwise methodologies including ontology-based text mining, self-organizing maps, reliability and cost optimization for identifying manufacturing faults, mapping them to design information and finally optimizing design parameters for maximum reliability and minimum cost respectively. The DSS analyses warranty databases which collect the warranty failure information from the customers in a textual format. To extract the hidden knowledge from this, an ontology-based text mining based approach is adopted. A data mining based approach using Self Organizing Maps (SOM) has been proposed to draw information from the warranty database and to relate it to "the manufacturing data. The dusters obtained using SOM are analyzed to identify the critical regions, i.e., sections of the map where maximum defects occur. Finally, to facilitate the correct implementation of design parameter changes, the frequency and type of defects analyzed from warranty data are used to identify areas where improvements have resulted in the greatest reliability for the lowest cost.