Detecting Social Media Bots with Variational AutoEncoder and k-Nearest Neighbor

Xiujuan Wang,Siwei Cao,Yutong Shi,Yi Sui,Qianqian Zheng,Kangfeng Zheng

doi:10.3390/app11125482

Abstract

Malicious social media bots are disseminators of malicious information on social networks and seriously affect information security and the network environment. Efficient and reliable classification of social media bots is crucial for detecting information manipulation in social networks. Aiming to correct the defects of high-cost labeling and unbalanced positive and negative samples in the existing methods of social media bot detection, and to reduce the training of abnormal samples in the model, we propose an anomaly detection framework based on a combination of a Variational AutoEncoder and an anomaly detection algorithm. The purpose is to use Variational AutoEncoder to automatically encode and decode sample features. The normal sample features are more similar to the initial features after decoding; however, there is a difference between the abnormal samples and the initial features. The decoding representation and the original features are combined, and then the anomaly detection method is used for detection. The results show that the area under the curve of the proposed model for identifying social media bots reaches 98% through the experiments on public datasets, which can effectively distinguish bots from common users and further verify the performance of the proposed model.

Highlights

With the explosive growth in social network services, they have become commonplace for communication and as a platform for building relationships [1]; most people are willing to record their lives and express their views on social media platforms
We selected the appropriate decoding features and fused the original features as the input of each algorithm in this part. These algorithms are very common in the field of anomaly detection, including the angle-based outlier detector (ABOD) [33], cluster-based local outlier factor (CBLOF), feature bagging, Histogram-based Outlier Score (HBOS), Isolation Forest (IForest), One-Class Support Vector Machines (OCSVM), fully connected autoencoder (AE), and Variational AutoEncoder (VAE)
We proposed a new social media bot detection method, using VAE to encode and decode features, and normal data for training

Summary

Introduction

With the explosive growth in social network services, they have become commonplace for communication and as a platform for building relationships [1]; most people are willing to record their lives and express their views on social media platforms. A social media bot is an abnormal user on a social media platform. Social media bots have increased exponentially in recent years and are essentially social media accounts that are completely or partially controlled by a computer algorithm. They can automatically generate content and interact with human users, usually disguised as humans [2]. They can create false accounts, steal user privacy, send spam, spread malicious links, and perform other activities. Social media bots have become the “cancer” of social networks [4]

Methods

Results

Conclusion