Abstract
Machine learning and data mining can extract desired knowledge or interesting patterns from existing databases and ease the development bottleneck in building expert systems or decision support systems. In the past, the rough-set theory has been widely used in dealing with data classification problems. Most conventional mining algorithms based on the rough-set theory identify relationships among data using crisp attribute values; however, data with quantitative values are commonly seen in real-world applications. This paper thus deals with the problem of producing a set of maximally general certain and possible rules from quantitative data. A new method that combines the rough-set theory and the fuzzy-set theory, is thus proposed to solve this problem. It first transforms each quantitative value into a fuzzy set of linguistic terms using membership functions and then calculates fuzzy lower approximations and fuzzy upper approximations, after which certain and possible rules are generated based on the fuzzy approximations. The proposed algorithm is shown to be a general approach to the traditional rough-set algorithm for crisp data. Thus, it can be used to process both crisp and quantitative data.