Compact Weighted Class Association Rule Mining using Information Gain

TL;DR

This paper introduces a compact weighted class association rule mining method that enhances classification accuracy by generating fewer, high-quality rules using a novel weight calculation based on the HITS model.

Contribution

It proposes a new associative classification algorithm that selects a key attribute and uses HITS-based weights to improve rule quality and classification performance.

Findings

Fewer high-quality rules are generated.

Classification accuracy is improved.

The method effectively incorporates semantic significance through weights.

Abstract

Weighted association rule mining reflects semantic significance of item by considering its weight. Classification constructs the classifier and predicts the new data instance. This paper proposes compact weighted class association rule mining method, which applies weighted association rule mining in the classification and constructs an efficient weighted associative classifier. This proposed associative classification algorithm chooses one non class informative attribute from dataset and all the weighted class association rules are generated based on that attribute. The weight of the item is considered as one of the parameter in generating the weighted class association rules. This proposed algorithm calculates the weight using the HITS model. Experimental results show that the proposed system generates less number of high quality rules which improves the classification accuracy.