This document discusses two algorithms for mining association rules from horizontally partitioned databases while preserving privacy. The first algorithm is a two-phased approach that uses encryption for mining large itemsets in the first phase and then introduces random numbers to preserve privacy in the second phase. The second algorithm uses a technique called CK Secure Sum that breaks each site's data into segments and changes neighbors between sites in each round to limit information leakage. Both algorithms aim to allow association rule mining across distributed datasets without revealing private information from individual sites.