Paper Info
Title
Comparative study of decision rule induction approaches involving rough sets for meningitis categories' discrimination
Abstract
In this paper and in order to design a smart decision support system able to discriminate the community-acquired meningitis categories, we studied three different decision rule induction approaches. Two of these approaches are based on rough sets but differ in the number of stages in their processes and in the method used to generate rules. The first approach contains three stages namely continuous attributes' discretization, conditional features' reduction and decision rule induction and it applies the Indiscernibility Based Rule (IBR) algorithm. However, the second only includes two stages, which are discretization and decision rule generation, and utilizes the LEM2 (Learning from Examples Module, version 2) method. The third approach is similar to the second one except that it determines decision rules relied on AQ (Algorithm Quasi-optimal) or CN2 algorithm. The experiments were conducted on two multidimensional databases, one derived from the other, of clinical records concerning 310 cases of bacterial or viral meningitis. The obtained results showed that: combined with the "Greedy Heuristic for Computing Decision Reducts and Approximate Decision Reducts" (GHCADR) and the "Equal Width Interval Discretization" (EWID) methods, IBR algorithm offers the best performances in term of prediction accuracy, quality and compactness of generated decision rule sets.
Year
DOI
Venue
2019
10.1145/3368756.3369066
Proceedings of the 4th International Conference on Smart City Applications
Keywords
DocType
ISBN
attribute reduction, classification, discretization, meningitis, rough sets, rule induction
Conference
978-1-4503-6289-4
Citations 
PageRank 
References 
0
0.34
0
Authors
4
Name
Order
Citations
PageRank
Abdelkhalek Hadrani100.34
Karim Guennoun2213.63
Rachid Saadane32211.94
Mohammed Wahbi400.34