With the rapid growth of network technology and data storage, big data applications have expanded significantly, making data mining essential for extracting valuable insights. However, protecting sensitive information during data mining is critical. This paper proposes two privacy-preserving methods for federated data mining. The first method uses pseudonymized SHA encryption and clustering to mine frequent itemsets while ensuring privacy. The second one employs the dummy itemset injection to enhance privacy by adding fake itemsets, reducing the risk of reverse engineering. Experiments are made to validate the effectiveness of the proposed methods, revealing the impact of character modifications on mining performance and offering a solid foundation for privacy-preserving federated mining.