CNN‐based malicious user detection in social networks

摘要

Following the advances in various smart devices, there are increasing numbers of users of socialrnnetwork services (SNS), which allows communication and information sharing in real time withoutrnlimitations on distance or space. Although personal information leakage can occur through SNS,rnwhere an individual's personal details or online activities are leaked, and various financial crimesrnsuch as phishing and smishing are also possible, there are currently no countermeasures. Consequently,rnmalicious activities are being conducted through messages toward the users who are inrnfollow or friend relationships on SNS. Therefore, in this paper, we propose a method of assessingrnfollow suggestions from users with less likelihood of committing malicious activities through anrninformation‐driven follow suggestion based on a categorical classification of interests using bothrnthe images and text of user posts. We ensure the objectiveness of interest categories by definingrnthese based on DMOZ, which is established by the Open Directory Project. The images and textrnare learnt using a convolutional neural network, which is one of the machine learning techniquesrndeveloped with a biological inspiration, and the interests are classified into categories. Users with arnlarge number of posts are defined as certified users, and a database of certified users isrnestablished. Users with similar interests are classified, and the similarity distances between certifiedrnusers and users are measured, and a follow suggestion is generated to the certified user withrnthe most similar interest. Using the method proposed in this paper to classify the interest categoriesrnof certified users and users, precisions of 80% and 79.8% were obtained, respectively, and thernoverall precision was 79.93%, indicating a good classification performance overall. It is expectedrnthat the method proposed in this paper can be used to provide follow suggestions of users withrnless likelihood of malicious activities based on the information posted by the user.