The paper introduces a new association rule mining algorithm that leverages an undirected item set graph to improve efficiency by scanning the database only once, as opposed to traditional algorithms like the Apriori algorithm which require multiple scans and generate numerous candidate item sets. The proposed method identifies frequent item sets and generates association rules based on the support and confidence metrics directly from the graph. Results demonstrate a significant reduction in space and time complexity when compared to existing mining techniques, making it suitable for managing large transactional databases.