Spam Tags Detection and Protection Using Tags' Relationship Based Anti-Spam Approach

With the emergence of social networking, people use to spend most of their time browsing the internet. Folksonomy the other name used for user vocabulary, which is in the form of a collection of tags, is a common feature of most of the social networks. It allows applying public tags to online items (images, music, videos etc.). In this respect, tags have gained a lot of interest in social networking. Tags are becoming popular for searching and labeling resources available on Internet (images, videos etc.). So, folksonomy assumes an imperative part to index data. However, at the same time, spammers also emerge on the internet globe and they start misleading and annoying users from their activities by assigning spam tags to resources. These spam tags cause waste of the user's precious time. In this article tags' relationship-based anti-spam approach is proposed and tested. It detects unwanted/spam tags. The approach we propose considers the relationships among tags. As folk tags set is always rich in relationships. Presence of at least one relationship between any two tags of a tag set will make it a legitimate one. The analysis of results proves the proposed approach is effective in providing protection against spam tags. In addition, it increases the understandabilityof the resource and brings precision in search and retrieval.