Abstract Collaborative filtering recommendation systems that utilize graph convolutional neural networks (GCNs) often emphasize on training speed and accuracy, while overlooking the critical issue of fairness. This paper introduces FairGCF, a novel graph-based collaborative filtering model designed to address fairness concerns in GCN-based recommendations. FairGCF improves the representational capacity of graph collaborative filtering models by identifying dependencies between nodes in the graph, resulting in fairer and more accurate recommendations. The model constructs a dependency graph between users and items to capture higher order interaction patterns, aggregating features from users and items with similar dependencies to enrich their representations. Additionally, a fairness loss factor is incorporated to minimize disparities between positive and negative samples, thereby promoting fairness. Experiments on three public datasets demonstrate that FairGCF improves the Recall metric by an average of 6%, NDCG by 8.79%, Precision by 3.68%, and F1 score by 3.62%. Notably, the model is carefully designed to avoid bias toward specific users or items.

Data privacy preservation and secure sharing are key technical challenges faced by smart wearable healthcare Internet of Things (IoT) systems. Blockchain technology enables privacy preservation for medical data through encryption. However, conventional data encryption hampers data analysis and sharing, and decrypted data still carries the risk of leakage. Homomorphic encryption is a technique that allows computation directly on encrypted data without decryption, thus reducing the risk of data leakage during sharing. In this article, we propose a blockchain-based privacy preservation and sharing scheme for healthcare IoT data. First, we use an improved homomorphic encryption technique to encrypt and process electronic health records (EHRs), optimizing the modular exponentiation process with a fast exponentiation algorithm, enabling users to efficiently perform data computation and analysis while keeping the data encrypted. Second, we employ symmetric searchable encryption (SSE) to encrypt homomorphic keys and user identity information, and use a Bloom filter as the mapping structure between data keywords and unique identifiers. This approach enhances search efficiency while preserving data privacy, allowing for secure search and analysis on ciphertext. Finally, smart contracts are designed to implement access control during the data-sharing process, increasing the security and transparency of data sharing. Experimental results show that the proposed homomorphic encryption scheme reduces the encryption and decryption time by an average of 34% under different key sizes, while the optimized SSE technique keeps ciphertext retrieval time at a constant level. The proposed scheme provides an effective solution for secure and efficient data analysis and retrieval, ensuring privacy preservation for the secure use and sharing of medical data.

Summary Hadoop is a typical framework for processing big data. Task scheduling algorithms have a significant impact on the processing performance of Hadoop clusters. Existing scheduling algorithms of Hadoop fail to consider the performance differences between nodes in heterogeneous Hadoop clusters, causing problems such as uneven task allocation and low resource utilization. Aiming to solve this problem, we propose a spider monkey optimization‐based scheduling algorithm (SMOSA) for heterogeneous Hadoop. First, the cluster heartbeat mechanism is used to obtain information such as memories and CPUs of nodes to comprehensively consider the actual load capacity of each node. Then, the spider monkey optimization algorithm is adopted to find the optimal mapping relationship between tasks and resources by taking the task completion time as the objective function and updating the position of the spider monkey. Finally, we calculate the remaining rate of node hardware resources, and according to the task type, the node with the higher remaining rate of resource will give priority to the task. Data are compressed for I/O type tasks to reduce disk operations and increase the speed of task execution. Experimental results show that, compared with existing scheduling algorithms, the SMOSA can effectively reduce task execution time and can significantly improve scheduling efficiency and task execution speed especially in heterogeneous Hadoop clusters. For different types of tasks, the execution time can be reduced by up to 19%.