ASSOCIATIVE CLASSIFIER BASED ON HIERARCHICAL CLUSTERING
Keywords:
Associative Classification, Class Association Rules (CARs), Hierarchical Clustering, Rule-Based Learning, Data Mining, Classification Accuracy, Knowledge Discovery, Interpretable Machine Learning, Apriori Algorithm, Adult datasetAbstract
This article proposes a novel associative classification approach called Associative Classifier based on Hierarchical Clustering (ACHC), which integrates association rule mining with hierarchical clustering techniques to improve classification accuracy and interpretability. The method utilizes the Apriori algorithm to extract class association rules from categorical data and clusters the rules based on their antecedent similarity using agglomerative clustering. The classification process involves matching the input instance to the most confident rule within its cluster. The proposed method is evaluated on the Car Evaluation dataset and shows promising performance in handling symbolic and rule-based knowledge discovery tasks, providing both classification and clustering insights. The article also presents experimental results and discusses the advantages of using rule clusters for efficient decision making.
