Which algorithm is designed to efficiently uncover frequent itemsets and patterns in large data sets?

• A. FP-Growth
• B. Apriori
• C. Eclat
• D. GSP

Answer: (A)

Efficiently uncovering frequent itemsets in large data sets is achieved with FP-Growth, which relies on a compact FP-tree structure to keep track of item frequencies and how they co-occur. After a first pass to count item frequencies and prune those that are infrequent, the remaining items are ordered by frequency and transactions are inserted into the FP-tree in that order. The resulting tree often compresses many transactions that share prefixes, dramatically reducing the amount of data that needs to be scanned again.

Mining proceeds directly on the tree by building conditional pattern bases for each frequent item and recursively mining conditional FP-trees to extend patterns. This avoids generating a large set of candidate itemsets, which is what makes the approach scalable to big data. In contrast, other methods tend to rely on generating many candidates or multiple full database scans (for example, candidate-based approaches can explode in size; some vertical representations or sequential-pattern algorithms have different trade-offs or are aimed at other types of patterns). FP-Growth’s combination of a compact data structure and candidate-free mining is what makes it particularly well-suited for large data sets.