It is necessary to solve the inaccurate data arising from data reliability ignored by most data fusion algorithms drawing upon collaborative filtering and fuzzy network theory. Therefore, a model is constructed based on the collaborative filtering algorithm and fuzzy network theory to calculate the node trust value as the weight of weighted data fusion. First, a FTWDF (Feedback Trust Weighted for Data Fusion) is proposed. Second, EEFA (Efficiency unequal Fuzzy clustering Algorithm ) is introduced into FTWDF considering the defects of the clustering structure caused by ignoring the randomness of node energy consumption and cluster head selection in the practical application of the existing data fusion algorithm. Besides, the fuzzy logic is applied to cluster head selection and node clustering. Finally, an FTWDF-EEFA clustering algorithm is constructed for generating candidate cluster head nodes, which is verified by simulation experiments. The comparative analysis reveals that the accuracy of the FTWDF-EEFA clustering algorithm is 4.1% higher than that of the TMDF (Trust Multiple attributes Decision-making-based data Fusion) algorithm, and 8.3% higher than that of LDTS ( Larger Data fusion based on node Trust evaluation in wireless Sensor networks) algorithm. It performs better in accuracy and recommendation results during the processing of ML100M dataset and NF5M dataset. Besides, the new clustering algorithm increases the survival time of nodes when analyzing the number of death nodes to prolong networks’ lifespan. It improves the survival period of nodes, balances the network load, and prolongs networks’ lifespan. Furthermore, the FTWDF-EEFA clustering algorithm can balance nodes’ energy consumption and effectively save nodes’ overall energy through analysis. Therefore, the optimized algorithm can increase the lifespan of network and improve the trust mechanism effectively. The performance of the algorithm has reached the expected effect, providing a reference for the practical application of the trust mechanism in networks.

中文翻译：

---

多媒体网络中反馈信任权重的信任机制

有必要利用协同过滤和模糊网络理论来解决大多数数据融合算法忽略的数据可靠性导致的数据不准确。因此，基于协同过滤算法和模糊网络理论构建模型，计算节点信任值作为加权数据融合的权重。首先，一个FTWDF（数据融合的反馈信任加权）被提议。第二，EEFA（效率不等模糊聚类算法）) 是考虑到现有数据融合算法在实际应用中忽略节点能耗和簇头选择的随机性造成的聚类结构的缺陷，引入FTWDF。此外，模糊逻辑被应用于簇头选择和节点聚类。最后，构建了一种FTWDF-EEFA聚类算法，用于生成候选簇头节点，并通过仿真实验进行了验证。对比分析表明，FTWDF-EEFA聚类算法的准确率比传统聚类算法高4.1%。TMDF（Trust Multiple attributes 基于决策的数据融合）算法，比算法高出 8.3%LDTS(无线传感器网络中基于节点信任评估的大数据融合）算法。在处理 ML100M 数据集和 NF5M 数据集时，它在准确度和推荐结果方面表现更好。此外，新的聚类算法在分析死亡节点的数量时增加了节点的生存时间，从而延长了网络的寿命。提高节点生存期，平衡网络负载，延长网络寿命。此外，FTWDF-EEFA聚类算法通过分析可以平衡节点的能量消耗，有效节省节点的整体能量。因此，优化后的算法可以有效地增加网络的寿命，改善信任机制。该算法的性能达到了预期效果，为信任机制在网络中的实际应用提供了参考。