PROJECT TITLE :

An Adaptive Social Spammer Detection Model with Semi-supervised Broad Learning

ABSTRACT:

Mobile Social Networks consist of a sizable number of members who pass on messages to one another in a collaborative manner. On the other hand, spammers will either post links to viruses and advertisements or follow a large number of users, both of which will result in a large number of misleading messages being spread throughout mobile social networks. In this paper, we propose a model for the adaptive social spammer detection, also known as ASSD. We construct a spammer classifier by making use of a limited number of labeled patterns in conjunction with a few unlabeled patterns. When compared to other traditional supervised learning methods, the accuracy of the prediction is quite high. The application of ASSD also cuts down on the amount of time and effort needed to determine the identities of members of a social group. Because social spammers frequently alter their behavior in an effort to fool the spammer detection model, an incremental learning method has been developed to update the spammer detection model in an adaptive manner, without requiring the model to first be retrained. The Social Honeypot Dataset is utilized in our analysis of ASSD so that we can evaluate it in comparison to other supervised and semi-supervised Machine Learning methods. The results of the experiments show that the proposed model performs significantly better than the baseline methods in terms of recall and precision. In addition, Automatic Social Media Surveillance and Defense (ASSD) keeps a high detection accuracy by dynamically updating the model with newly generated social media data.