Improved Algorithm of FP-Growth Based on Strong Data Correlation

The FP-Growth algorithm takes a large proportion of the time spent on mining frequent itemsets and building FP-tree. This paper proposes a method based on the original FP-Growth algorithm, which sorts the strongly correlated data in advance, that is, traverses the filtered data set one more time, searches for something that has relationships in the database in advance, merge collections with the sets with strong association rules to reduce the number of child nodes of FP-Tree, thereby speeding up the generation of FP-Tree, which reduces the memory occupied by FP-Tree and facilitates the search of following items. Experiments show that the improved FP-Growth algorithm is better than the traditional FP-Growth algorithm in both the running time spent on frequent itemsets and the total running time of the algorithm.