Detection Of Criticism and Hatespeech Text Formulation on Online Sosial Network Twitter For Semantic Recommendation System Framework Migunani (a*), Adi Setiawan (b), Irwan Sembiring (c)
Science and Computer Technology University
Satya Wacana Christian University
Abstract
User opinions on high-volume social media and various themes provide relevant information for sentiment analysis. This information can be collected and analyzed using a natural language processing with a monitoring system to support classification of criticism and hate speech. Regarding monitoring results, a knowledge-based recommendation system with sentiment analysis is supported to send messages to user in order to use positive sentences are not offensive, polite, wise and motivational for users with hateful attitudes. Detection of sentences containing criticism and hate speech using Bag Of Word and Convolutional Neural Network to detect hate speech and criticism sentence via Tweeter. The detection results are used for the semantic recommendation system framework that includes sentiment analysis and predicts hate speech through similarities between previous cases by finding the value of proximity or closeness between text to be a measurement baseline to criticism and hate speech clasification.
Keywords: Social Media, Hate Speech, Bage of Word, Convolutional Neural Network, Recommendation System
