Cost-Based Heterogeneous Learning Framework for Real-Time Spam Detection in Social Networks With Expert Decisions

Abstract

With the widespread use of social networks, spam messages against them have become a major issue. Spam detection methods can be broadly divided into expert-based and machine learning-based detection methods. When experts participate in spam detection, the detection accuracy is fairly high. However, this method is highly time-consuming and expensive. Conversely, methods using machine learning have the advantage of automation, but their accuracy is relatively low. This paper proposes a spam-detection framework that combines and fully exploits the advantages of both methods. To reduce the workload of the experts, all messages are first analyzed via a primary machine learning filter, and those that are determined to be normal messages are allowed through, whereas suspicious messages are flagged. The flagged messages are subsequently analyzed by an expert to enhance the overall system accuracy. In the filtering process, costbased machine learning is used to prevent the fatal error of misidentifying a spam message as a normal message. In addition, to obviate the continuously evolving spam trends, a module that periodically updates the expert-diagnosis results on the training dataset is incorporated into the framework. The results of experiments conducted, on an imbalanced dataset of spam tweets and normal tweets in a ratio similar to the actual situation in real life, indicate that the proposed framework has a spam-detection rate of almost 92.8%, which is higher than that of the conventional machine learning technique. Furthermore, the proposed framework delivered stable high performance even in an environment where social network messages changed continuously, unlike the conventional technique, which exhibited large performance deviations.INDEX TERMS Expert decision making, machine learning, real-time spam detection, social network, Twitter spam.

Achieving Online and Scalable Information Integrity by Harnessing Social Spam Correlations

“…Each Spiral group consists of a root, branches, and many leaf agents. The root deploys the spam detection model (for example, using Random Forest [39,41] or Naive Bayes [39,69]) based on the training dataset from Phase 1. This model is used in the local data processing in a single group (details in subsection IV-E) and global data processing in many groups (more details can be found in subsection IV-F).…”

A comprehensive survey on online social networks security and privacy issues: Threats, machine learning‐based solutions, and open challenges

“…In their scheme, 153 the authors used evolutionary algorithms and ensemble classification methods to diagnose OSN spams. Choi and Jeon 154 proposed a spam-detection framework where all messages are filtered using ML methods and suspicious messages are further examined by experts for detecting spam in social networks [155][156][157][158] and many more ML-based spam detection approach have been proposed in last few years.…”

Scalable Learning Framework for Detecting New Types of Twitter Spam with Misuse and Anomaly Detection

“…However, honeypot-based methods suffer from several issues, such as poor adaptability and scaling, high volume of data processing, and poor portability. Their implementation is also time-consuming as they require expert intervention [2,12]. To overcome such drawbacks of traditional methods, recent studies have employed and developed machine learning (ML) and deep learning (DL)-based methods to detect spam and malicious behavior on social media.…”

“…When misuse detection is used, excellent performance is achieved in detecting known attacks because previously known data are learned and subsequently used for spam detection. However, misuse detection methods are ineffective against previously not learned spam attacks evolving in real time, and their major disadvantage is performance degradation due to class imbalance caused by imbalanced amounts of spam and non-spam data [12,14]. In particular, spammers have collaborated and rapidly shifted their attack strategies in response to recent defense strategies.…”