An efficient approach to reduce rules for fault diagnosis based on rough set

Analysis of wear data of vehicle's transmission system is often concerned with treatment of incomplete knowledge. Existing techniques of data analysis are mainly based on quite strong assumptions, are unable to deduce conclusions from incomplete knowledge or inconsistent data of information. The rough set with the lower and upper approximation of the concept is a useful notion for the classification of objects when the available information is not adequate to represent classes using precise sets. In this paper, the Rough set theory is deeply investigated, and an approach reduction of attributes and inducing rules based on rough set theory in a fault diagnosis system is proposed. Rules induced from the lower approximation of the class certainly describe the class (certain rules), in the other hand, rules induced from the upper approximation of the class describe only possibly case (possible rules). This approach is applied for analysis of emission spectrum data of lubricating oil of vehicle's transmission system. There are 14 conditional attributes and 1 decisional attribute in the information system table. The rule-base are constituted with certain and possible rules of analysis.