Improved Denclue outlier detection algorithm with differential privacy and attribute fuzzy priority relation ordering

Outlier detection is an important method in data mining. Although Denclue algorithm particularly good at finding clusters of arbitrary shape and detecting outliers, it does not protect the user's privacy well operation process. In this paper, differential technology introduced into to ensure security application outlier detection. Firstly, used add Laplacian noise density realize sensitive information hiding between objects. Secondly, order compensate for decrease accuracy caused by noise, entropy weight distance was amplify influence attributes algorithm, function calculate each point. Finally, through ordering fuzzy priority relation, a new measure index defined analogy degree outliers attributes. According index, are reordered improved. A based on attribute (EAF-DP-Denclue) proposed. The numerical results experiment show that performance EAF-DP-Denclue more than traditional algorithms, identification process protects information, better DP-DBScan algorithm.