Partial Dominance: A New Framework for Top-k Dominating Queries on Highly Incomplete Data
Faruk Hossen, K. M. Azharul Hasan, H. M. Abdul Fattah, Tatsuo Tsuji · 2023
The Top-k dominance locates the top k objects that happen to be most dominant across a dataset. It makes it possible for data analysts to unearth a previously unknown pattern known as dominant objects, which may be put to use in a variety of applications including decision support systems, suggested system, etc. Hence, TKD becomes an important decision making tool for many organizations. To dominate a record over another record, it is necessary for the the record to be dominated by all the attributes in the dataset. But it can happen that one entity A dominates most of the attributes of another entity B and B dominates A by a very small number of attributes. In this circumstance it can not be said that A dominates B or B dominates A using traditional dominating score computation scheme. When none dominates, traditional top-k dominating query cannot find top k objects for such datasets. This study proposes a framework to deal with such datasets that computes top k dominating query more accurately and more realistic manner form any kind of dataset. In order to make the solution realistic, this study incorporates the concept of partial dominance that can calculate the dominating objects from such dataset. The non-candidate objects are eliminated using the concept of data bucketing. The experimental result shows that the proposed model outperforms existing important approaches for top-k dominant query.