User Audit Model Based on Attribute Measurement and Similarity Measurement

Xiaohui Yang,Ying Sun

doi:10.1155/2020/8387672

Abstract

The Internet of Things (IoT) is an open network. And, there are a large number of malicious nodes in the network. These malicious nodes may tamper with the correct data and pass them to other nodes. The normal nodes will use the wrong data for information dissemination due to a lack of ability to verify the correctness of the messages received, resulting in the dissemination of false information on medical, social, and other networks. Auditing user attributes and behavior information to identify malicious user nodes is an important way to secure networks. In response to the user nodes audit problem, a user audit model based on attribute measurement and similarity measurement (AM-SM-UAM) is proposed. Firstly, the user attribute measurement algorithm is constructed, using a hierarchical decision model to construct a judgment matrix to analyze user attribute data. Secondly, the blog similarity measurement algorithm is constructed, evaluating the similarity of blog posts published by different users based on the improved Levenshtein distance. Finally, a user audit model based on a security degree is built, and malicious users are defined by security thresholds. Experimental results show that this model can comprehensively analyze the attribute and behavior data of users and have more accurate and stable performance in the practical application of the network platforms.

Highlights

Introduction eInternet of ings (IoT) is the latest evolution of the Internet, including a great deal of connected physical devices and applications [1]
The Internet of ings to hardware has social attributes. erefore, to maintain the security of the Internet of things and identify malicious users in the network, in response to the above problems, a user audit model based on attribute measurement and similarity measurement (AMSM-UAM) is proposed by taking the social platform of microblog with a large user volume as an example
According to the attribute vectors corresponding to the five user features of microblog level Al, big-V certification AV, personal information integrity Ap, number of followers Af, and number of fans As, and combined with the weight vector β, the user attributes are numerically represented to reflect the user’s own security degree, as in the following equation: Attribute measurement (AM)(u) 􏼐Al, Av, Ap, Af, As􏼑 · (β)T

Summary

Related Work

Malicious user identification methods based on abnormal behavior detection have attracted considerable attention. Erefore, to maintain the security of the Internet of things and identify malicious users in the network, in response to the above problems, a user audit model based on attribute measurement and similarity measurement (AMSM-UAM) is proposed by taking the social platform of microblog with a large user volume as an example. 3. Model Description e key to the construction of the AM-SM-UAM is to rationally quantify the user’s attribute information and behavior information, to realize the identification of malicious users and to ensure the smooth operation of the microblog. The similarity measurement algorithm is constructed to process users’ blog information and evaluate the similarity of users with different attribute values in blog keywords, so as to achieve the purpose of computing the similarity of textual data. The similarity measurement algorithm is constructed to process users’ blog information and evaluate the similarity of users with different attribute values in blog keywords, so as to achieve the purpose of computing the similarity of textual data. e user’s attribute information and blog text information are considered comprehensively from the two aspects of user attribute and spontaneous behavior to obtain user security degree

Model Construction

Experiments

G2 G3 G4 G5 G6 G7 G8 G9 G10 Experimental group