A new approach to estimate the missing value based on neighbor granular domain in the incomplete information system

It is an important research subject to estimate the missing value in the incomplete information system. In this paper, we formally define the incomplete information system, incomplete equivalence class, reflexive neighborhood, granularity, granularity domain, rough granularity domain, neighborhood granularity domain, and the property of neighborhood granularity domain. Then we put forward a new method based on neighborhood granularity domain to estimate missing value of the incomplete information system. First calculate the object incomplete equivalence class of the condition attributes and decision attributes, and then obtain granularity domain elements from the incomplete equivalence class, and calculate the rough granularity domain and the neighborhood granularity domain, and then divide all calculated objects of neighborhood granularity domain into according equivalence class, finally delete other equivalence class of the object. We presents the algorithm and examples, and the experimental results show that the algorithm is fast and the missing value estimation accuracy is high.