JOINT: Join Optimization and Inference via Network Traversal

Traditional relational databases require users to manually specify join keysand assume exact matches between column names and values. In practice, thislimits joinability across fragmented or inconsistently named tables. We proposea fuzzy join framework that automatically identifies joinable column pairs andtraverses indirect (multi-hop) join paths across multiple databases. Our methodcombines column name similarity with row-level fuzzy value overlap, computesedge weights using negative log-transformed Jaccard scores, and performs joinpath discovery via graph traversal. Experiments on synthetic healthcare-styledatabases demonstrate the system's ability to recover valid joins despitefuzzified column names and partial value mismatches. This research has directapplications in data integration.