A Framework to Detect Crime through Twitter Data in Cyberbullying with EFFDT Model

Abstract

Due to enormous availability of internet, social networking and micro blogging websites such as twitter, instagram, are increased. The users registered with the website utilize the webpage as a platform to express their thoughts as comments and convey their opinions to the global users. Cyber bulling has become one of the serious issues in recent days because of controversial comments and thoughts exposed in the micro blogging websites that impacts the society and certain group of peoples in a negative way. The amount of negatively impacted comments on micro blogs are getting increased in recent days, it is required to identify those traits as a crime impacted feeds for police attention. Texting is a technique used to make a classification of large weights into clusters of related data. The proposed framework is focused on collecting various tweets from the micro blogs, further pre-processing it using natural language processing (NLP) for the features selection, is implemented using partial spam Optimization (PSO). Based on the feature extraction process, the classification model is developed using ensemble approach. The proposed approach considers Ensemble feed forward decision tree (EFFDT) model to classify different types of negative tweets from the given database. The machine learning algorithm namely Support vector Machine (SVM) algorithm and K-Nearest Neighbour (KNN) algorithm are used for comparison with the proposed method. The performance result of these algorithms are compared in terms of precision, F1Score, accuracy and further compared to the state of art approaches.