Intuitionistic fuzzy rough sets and fruit fly algorithm for association rule mining

Association rule mining (ARM) is a data mining technique for identifying frequently occurring item groupings in transactional datasets. The frequent item recognition and ARM development are two critical processes in ARM. Association rules are generated using minimum support and confidence metrics. Numerous methods have been projected by scholars for the purpose of generating association rules. In general, a large number of datasets can be evaluated, necessitating an increased number of database searches. Additionally, data analysis may not require all of the characteristics of the input data. The suggested association rule mining project is conducted on seven biological data sets from the University of California, Irvine (UCI). As a result, the initial part of this study endeavour employs a dimensionality reduction method that significantly shrinks the size of the data collection. The suggested approach efficiently finds the database's significant properties. To improve classification performance, the proposed approach eliminates extraneous features from the UCI database. The projected technique for dimensionality reduction is compared to intersection set theory extended frequent pattern and Dimensionality Reduction Using Frequency counT. The second stage recommends using an intuitionistic fuzzy-rough set (IFRS) in conjunction with the Fruit fly Algorithm (FFA) to identify common items and generate association rules. The suggested algorithm's efficiency is associated to particle swarm optimization and genetic algorithms that are built in accordance with IFRS. Experiments demonstrated that the recommended strategies achieved satisfactory results.The proposed IFRS-FFA method achieved 98.7% of recall, 98.5% of precision and 80.42% of accuracy on Vertebral of 3 class dataset.