A Research Effort to Categorize Social Posts Applying Two Phase Natural Language Processing Methodology
Geetanjali R. Salunke, Siddhasen Patil, Debnath Bhattacharyya, Hye-Jin Kim · International Journal of Advanced Science and Technology · 2017
Today with Advancement in network and Web Technology ratio of internet users is counting up.In current scenario, users are so much comfortable to Social Networks.Different OSN (Online Social Network) are used to connect with each other.Facebook is an emerged as vital source of communication.But security has been overlooked.Most of the facebook user spread and sent various type of content such as post messages, images, URLs, News about distinct topics, But major problem with facebook and other OSN a large amount of adult content are spread very rapidly through internet.Adult content are in the form of messages, URLs (Links containing adult web pages), images.Through internet people can easily spread adult content on facebook and OSN.There is need to provide the security to online social networks from adult content spread through web.We are proposing an effective solution for facebook content classification into adult and non adult.In this paper we are proposing an application which is used to classify facebook content into adult and non adult content at text and URLs level.To classify the facebook content into adult and non adult content we use natural language processing and machine learning techniques.We then categorize the classified facebook content into different topics such as business and industrial, law, government and politics, news and others.In addition, we enlarge our system for facebook data classification based on facebook post and comment posted on own wall and graphically represent overall classification of adult and non adult posts and categorization of user post.