You have done a interesting research. What are heuristics here? What problems does your solution solve in the real environment? Thank you.
Thank you for your remarks. Heuristic approach based on solving problem with good solution along with extensive experimental evaluations. In other words we can say that heuristic approach provides a fair solution in a shorter execution time. In this paper, we have proposed heuristic rules to compute approximations without going through the lengthy computational steps of traditional DRSA algorithm and we achieved considerable reduction in execution time. Our rules based on the definitions of preliminaries of dominance based rough set theory and properties of approximations. We avoid lengthy comparisons and repetitions by using these rules which ultimately reduce computational cost in terms of time. We have performed experiments on UCI datasets, extensively, to judge the performance of proposed heuristic approach, in comparison with static and non-static techniques.
By using this approach, we consequently reducing computational cost of all subsequent tasks based on approximations i.e., decision rule generations, reduct generations, classification, prediction etc. Our approach proposed a mechanism to update approximations in dynamic environment, much more efficiently. For example, in real environment, we can use proposed method to update the information regarding bankrupt chances of companies by analyzing preference ordered data in dynamic environment and for many other multi-criteria classification applications in real time environment.
