A New Algorithm in Association Mining, Amoeba for Finding Frequent Patterns Using Functional Dependency and Probability

In Data Mining, research on association mining has expanded widely in many application areas. APRIORI and FP-Growth are prominent algorithms in association mining for finding frequent patterns. These algorithms have their own shortcomings like space complexity and time complexity. Moreover constructed transaction data set in obligatory for these algorithms as input. Improved versions of these approaches have reduced the prior mentioned shortcomings. A new algorithm, Amoeba, was proposed to find chain of possible frequent patterns. This algorithm excludes construction of transaction dataset and calculating thresholds and includes probable occurrences of attribute values using functional dependencies.