Abstract:
Effective data analysis of big datasets is a challenge and when data changes over time,
new feature values are added, as a result, its complexity increases. An advanced approach of
rough sets theory known as the Dominance-based Rough Set Approach (DRSA) is a powerful
mathematical tool for identifying meaningful information in preference-ordered datasets. Data
analysis using DRSA is primarily based on the lower and upper approximations calculation,
which are very expensive to compute. It mainly used resources i.e., time execution and memory
consumption. When the data changes over time, approximation sets must be recalculated. As a
result, calculations that are repeating, increase the cost of approximation’s computation in the
real-time domain. The proposed approach computes approximations for increasing object values
with time. Results were compared with the conventional approach using UCI publicly available
datasets. When compared to the usual method, our suggested approach updated approximations
in less time, with an average reduction of 99.2%.
