Building a Competitive Associative Classifier

Sood, Nitakshi; Zaı̈ane, Osmar R.

doi:10.1007/978-3-030-59065-9_18

Cited by 5 publications

(2 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Confidence and support were used in the case of CBA and CMAR because the measures define reliability and generality of rules respectively. They were also commonly used in other studies [23,30,35,36] with other pre-defined measures for the same purpose. According to the surveys of Geng et al [16] and Sharma et al [33], there are more than 30 interestingness measures that can be employed to prioritize rules.…”

Section: Related Workmentioning

confidence: 99%

“…They proposed to split data into partitions, learning an independent model for each partition and finally merging them into a single model. Rule-pruning techniques were also considered in order to retain only statistically significant rules, such as the minimum χ 2 threshold in Venturini et al [36] and Fisher's exact test in Sood et al [35]. Emerging pattern based classifiers [11,12,14,19] are also rule-list based methods.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

MoMAC: Multi-objective optimization to combine multiple association rules into an interpretable classification

2021

View full text Add to dashboard Cite

A crucial characteristic of machine learning models in various domains (such as medical diagnosis, financial analysis, or real-time process monitoring) is the interpretability. The interpretation supports humans in understanding the meaning behind every single prediction made by the machine, and enables the user to assess trustworthiness before acting on the predictions. This article presents our work in building an interpretable classification model based on association rule mining and multi-objective optimization. The classification model itself is a rule list, making a single prediction based on multiple rules. The rule list consists of IF ... THEN statements that are understandable to humans. We choose these rules from a large set of pre-mined rules according to an interestingness measure which is formulated as a function of basic probabilities related to the rules. We learned the interestingness measure through multi-objective optimization, concentrating on two objectives: the classifier's size in terms of number of rules and prediction accuracy. The model is called MoMAC, "Multi-Objective optimization to combine Multiple Association rules into an interpretable Classification". The experimental results on benchmark datasets demonstrate that MoMAC outperforms other existing rule-based classification methods in terms of classification accuracy.

show abstract