Incremental knowledge discovering in interval-valued decision information system with the dynamic data

With the development of information technology and the rapid updating of data sets, the object sets in an information system may evolve in time when new information arrives and redundancy information leaves in real life. Interval-valued decision information systems are important type of data decision tables and generalized models of single-valued information systems. Fast updating the lower and upper approximations is the core technology of knowledge discovery that is based rough set theory in dynamic data environment. Consequently, in this paper we focus on incremental approaches updating approximations with dynamic data sets in interval-valued decision system. Firstly, we define an interval similarity degree by which a binary relation can be constructed, followed a rough set model be established. Then, incremental approaches for updating approximations are proposed and the incremental algorithms are shown. At last, comparative experiments on several UCI data sets show the proposed incremental updating methods are efficient and effective for dynamic data sets, namely, these approaches significantly outperform the classical methods with a dramatic reduction in the computational time.