基于模糊多目标决策的物联网大数据聚类算法

Big Data Clustering Algorithm of Internet of Things Based on Fuzzy Multi-objective Decision

现有的物联网大数据聚类算法容易受到相似性攻击,聚类效果较差。为了提升自适应能力,提出了一种基于模糊多目标决策的物联网大数据聚类算法。选取梯度下降法进行重复迭代,得到物联网事件的模糊置信度和支持度阈值,利用模糊C均值聚类算法获取最优模糊划分矩阵;建立目标决策矩阵,确定目标权重,明确理想决策目标和负理想决策目标,获取最终决策结果,从而实现物联网大数据的有效聚类。选取某电力企业的物联网大数据平台进行聚类实验,结果表明,该算法可有效聚类物联网平台中的海量数据,聚类结果的簇间区分度、簇间关联性和聚类敏捷性高。

The existing big data clustering algorithms of Internet of Things(IoT)are vulnerable to similarity attacks,and the clustering effect is poor.In order to improve the adaptive ability,an IoT big data clustering algorithm based on fuzzy multi-objective decision-making is proposed.Gradient descent method was selected for repeated iteration to obtain the fuzzy confidence and support threshold of IoT events,and obtain the optimal fuzzy partition matrix by fuzzy C-means clustering algorithm;The target decision matrix was established,the target weight was determined,the ideal decision target and the negative ideal decision target were defined,and the final decision result was obtained so as to realize the effective clustering of IoT big data.A clustering experiment is conducted on the IoT big data platform of an electric power enterprise.The results show that the algorithm can effectively cluster the massive data in the IoT platform;The clustering results have high inter-cluster discrimination and correlation,and have high clustering agility.