Dynamic computing rough approximations approach to time-evolving information granule interval-valued ordered information system

With the advent of Big Data era has seen both the volumes and update rates of data increase rapidly. The granular structure of an information system is evolving with time when redundancy data leaves and new data arrives. In order to quickly achieve the rough approximations of dynamic attribute set interval-valued ordered information system that the attribute set varies over time. In this study, we proposed two dynamic computing rough approximations approaches for time-evolving information granule interval-valued ordered information system which induced by the deletion or addition some attributes, respectively. The updating mechanisms enable obtaining additional knowledge from the varied data without forgetting the prior knowledge. According to these established computing rules, two corresponding dynamic computing algorithms are designed and some examples are illustrated to explain updating principles and show computing process. Furthermore, a series of experiments were conducted to evaluate the computational efficiency of the studied updating mechanisms based on several UCI datasets. The experimental results clearly indicate that these methods significantly outperform the traditional approaches with a dramatic reduction in the computational efficiency to update the rough approximations.