The aim of this project is to develop a solution for generating non-redundant association rules based on closed frequent itemsets. Association rule mining is a widely used technique in data mining and machine learning that aims to discover interesting relationships and patterns within large datasets. By identifying frequently co-occurring items, association rule mining can provide valuable insights and support decision-making processes in various domains.

However, traditional association rule mining approaches often generate a vast number of rules, many of which are redundant or irrelevant. Redundant rules can overwhelm users, making it difficult to extract meaningful and actionable knowledge from the results. This project tackles this challenge by leveraging the Charm algorithm, a powerful method for mining closed frequent itemsets.

The Charm algorithm is an efficient and effective approach for mining closed frequent itemsets, which are subsets of items that have the property of being "closed" under the given support threshold. Closed frequent itemsets capture all the maximal frequent patterns in the dataset, avoiding redundancy and providing a concise representation of the underlying associations.

In this project, we will implement the Charm algorithm and extend it to generate non-redundant association rules from the closed frequent itemsets. By eliminating redundant rules, we can improve the interpretability and usefulness of the mining results, enabling users to focus on the most interesting and informative patterns.

• Implementing the Charm algorithm for mining closed frequent itemsets.
• Extending the algorithm to generate non-redundant association rules.
• Developing efficient data structures and algorithms to handle large-scale datasets.
• Providing an intuitive and user-friendly interface for interacting with the mining results.
• Conducting thorough experiments and evaluations to assess the performance and effectiveness of the proposed approach.

http://fimi.uantwerpen.be/data/accidents.pdf