A frequent itemsets mining algorithm based on spatial partition

Established a complete lattices description for the problem of mining association rules, gave the lower limit of the problem scale, and put forward a spatial partition search based itemsets frequency calculation model. Based on the improved FP-tree, gave a frequent itemset mining algorithm UPM (upward partition mine) and proved that its complexity has achieved the minimum size of the problem. Performance experiments show that, compared with FP-Growth algorithm, UPM has an excellent performance in space and time.