An Improved Association Rule Mining Technique Using Transposed Database

Discovering the association rules among the large databases is the most important feature of data mining. Many algorithms had been introduced by various researchers for finding association rules. Among these algorithms, the FP-growth method is the most proficient. It mines the frequent item set without candidate set generation. The setbacks of FP growth are, it requires two scans of overall database and it uses large number of conditional FP tree to generate frequent itemsets. To overcome these limitations a new approach has been proposed by the name TransTrie, it will use the reduced sorted transposed database. After this it will scan the database and generate a TRIE, in the same step it will also compute the occurrences of each item. Then, using Depth first traversal it will identify the maximal itemsets, from which all frequent itemsets are derived using apriori property. It also counts the support of frequent itemsets which are used to find the valuable association rules