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

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 understanding meaning behind every single prediction made by machine, and enables user to assess trustworthiness before acting on predictions. This article presents our work building an interpretable classification model based association rule mining multi-objective optimization. itself a list, making multiple rules. list consists If ... THEN statements that are understandable humans. We choose these rules from large set pre-mined according interestingness measure which formulated function basic probabilities related learned through optimization, concentrating two objectives: classifier’s size terms number accuracy. called MoMAC, “Multi-Objective optimization combine Multiple Association into Classification”. experimental results benchmark datasets demonstrate MoMAC outperforms other existing rule-based methods