A practical solution for the classification in interval-valued information systems

A practical solution for the classification in interval-valued information systems Article

Caballero, A, Yen, K, Fang, Y. (2010). A practical solution for the classification in interval-valued information systems . 5(9), 735-744.

Complex information systems have to manage enormous quantity of data, representing different measurements or attributes from various sources. Classification of this set of data is a difficult task that frequently requires special methods for being solved. The problem of classification in information systems has been studied by many authors, and different methods have been developed. The use of rough sets, fuzzy logic, neural networks, entropy, or the combination of these methods has been used widely. When information is diffuse and the number of obtained values for each attribute is large, so is the number of rules obtained for the classification. Even worst is hidden information in the data that makes the process complicated. Due to this fact, an interval of values is defined for each attribute, moving from the minimum to the maximum obtained values in the collected database. This is what is defined as interval-valued information systems. Differently from other works, the concept of information measure is used in this paper, together with a fuzzy logic discrimination tool. Using these concepts, initially an attribute reduction is obtained and then fuzzy logic is applied for discriminating among the possible solutions. This method is simpler than others, and provides accuracy not less than the usually employed methods.