Detecting Rumor in Social Networks Using SVM and Comparing with Supervised Techniques

Abstract

A support vector machine (SVM) is a machine-learning algorithm that analyzes data for classification and regression analysis. SVM is a supervised learning method that looks at data and sorts it into one of the two categories. Using SVM in this project, The main aim is to categorize a message into rumor and non-rumor content which will increase the efficiency for identifying the rumor propagation in the social network. This rumor detection approach is based on the hashtags (#tags) classification and comparing different supervised learning techniques for getting better and accurate detection of rumors. The process of this approach is divided into four parts: Preprocessing, Feature Extraction, Sentiment Analysis and Classification. In this process, feature extraction methods used are Bow and TF-IDF .This methods intension is look at the histogram of the words with in the text .i.e. considering each word count as a feature .There are many classifiers available for detecting rumors. In This Proposed Model KNN (k-nearest), Gradient Boosting and Support Vector Machine achieves a high accuracy of SVM classifier than the other two. So that our proposed method gives the better accuracy for SVM classifier. The accuracy rates are    70.2% for KNN ,74% for Gradient Boosting and 75.3 %for SVM.In this project, the future work  will be extend as propagation of messages i.e. tracing of rumors and automatic blocking of that rumors.