A novel lightweight medical blockchain data query scheme

To address the limitations of lightweight medical blockchains in terms of data query efficiency and nonexistence proofs, we propose a novel lightweight medical blockchain data query scheme. First, the XGBoost algorithm is employed to predict medical data weights, with high-weight data storage near the blockchain's root nodes being prioritized, thereby optimizing the storage architecture and enhancing query efficiency. Second, an efficient query method that combines aggregated Bloom filters and Merkle-Huffman (MH) trees is designed. Through segmented filtering and weight optimization, the query path length is reduced, improving the on-chain data query performance. Finally, to address the challenge of the data nonexistence proof, we propose a multi-node collaborative verification mechanism that integrates Bloom filters with a dynamic reputation system. By adaptively selecting high-credibility nodes and employing multi-node consensus, false positives are minimized, ensuring query accuracy and reliability. Theoretical analysis and simulation results show that, compared with existing schemes, the proposed approach improves the query efficiency by approximately 15%. Moreover, integrating multi-node collaborative verification with a dynamic reputation mechanism effectively mitigates malicious attack risk and enhances system security, making it particularly suitable for resource-constrained scenarios such as mobile health care.