![]() | Only 14 pages are availabe for public view |
Abstract In this work Coronary Heart Disease KDD is treated based on two innovative hybrid models. First model is a combination of rough sets and flow graph. Rough sets theory is used in data dicretization, data reduction, decreasing data inconsistency and rule induction. Moreover, based on the concept of dynamic reduct; Rough sets is used in measuring the associated strength, certainty and coverage factors for each rule. A flow graph is generated based on rule associated factors .A probabilistic decision flow graph is configured based on rough set measurements of strength, certainty and coverage factors to describe the probabilistic decisions in purpose of classifying unseen objects. The new hybrid model helps decision-maker to conclude the rule with the highest priority. The other hybrid model “a combination system of rough sets and the machine learning algorithm C4.5 “is constructed. Rough set is used in data dicretization for continuous attribute values, data reduction and rule induction. Then Decision Tree is built based on the machine learning algorithm C4.5 where the most significant rules are generated. In contrast with the other previous models, a comparative study with previous models is illustrated to determine the best results based on classification accuracy. The details and limitations of the new hybrid models are discussed and future works are suggested. |