Discovering Conditional Functional Dependencies

Conditional Functional Dependencies are an extension of Functional Dependencies by supporting patterns ofsemantically associated constants, and can be used as rules for cleaning relational data. However, finding CFDs is undoubtedly an expensive process that involves intensive manual effort. To effectively identify data cleaning rules, we take 4 techniques for cleaning the data from sample relations. CFD Miner, is based on techniques for mining closed item sets, and is actually used to detect constant CFDs, namely, CFDs with constant patterns only. It offers a heuristicefficient algorithm for discovering patterns from a fixed FD. It leverages closed-item set mining to scale back search space. CTANE works well in the event of attributes of the relation is small and of course the support threshold is high, nevertheless it scales poorly whenever the attributes of the relation increases. Fast CFD is more efficient when the relation is larger than normal. Greedy Method formally driven by desirable properties of support and confidence. It studying the computational complexity of automatic series of optimal tables and providing an efficient approximation algorithm. Proposed algorithms discover both time and spacecomplexity of each algorithm to know which technique will just be helpful in this instance.