Extraction de règles d'association selon le couple support-MGK : Graphes implicatifs et Applications en didactique des mathématiques
Parfait Bemarisika · HAL (Le Centre pour la Communication Scientifique Directe) · 2016
In this work, we investigate the problem of mining positive and negative association rules and its applications in mathematics education. The extraction of this knowlege type is divided into two phases that are mining frequent patterns and generation of association rules from the set frequent patterns. Very often, the cost of mining frequent patterns in large and dense contexts is exponential, and the number of association rules generated could be excessively high, most of being not interesting. We propose a new model for mining frequent patterns, and a new model for generating the interesting association rules, using a new couple support-MGK. In the elaboration of implicatives graphs, most of existing methods use one quality measure, implication intensity, based on a gaussian approximation, which is not sef of lossing information. To resolve this defect, we proposed a new method using another quality measure, MGK. We propose a new algorithm of this implicatives graphs. The experiments conducted in data reference, show that our models reduce the costs of inputs/outputs and memory space. The application of these approach in mathematics education has highlighted the pratical value of our approach. We propose a new tool CHIC-MGK to serve as support for the educational searching activity. The effectiveness of this tool has been shown a real problem of education statistic, involving difficulties of our students in L1 level through proposed exercise solution.