Recognition of Disparaging Phrases in Social Media

Abstract

AbstractSocial media provides a space for individuals, where they can share their view or opinion very easily. The use of social media is growing at a rapid pace. Users can communicate more swiftly through social media platforms like Twitter, Facebook, and YouTube. The content of social media texts is generally made up of code-mixed comments/posts and replies, and it may contain harmful or non- offensive words. Sentiment analysis on social media provides businesses with a quick and effective approach to monitor public opinion on their brand, business, goods, and other topics. In recent years, a variety of features and approaches for training sentiment classifiers for fetched datasets have been investigated, with mixed results. Twitter is a popular social media platform. It provides businesses with a quick and efficient way to assess customers’ viewpoints on issues that are crucial to their performance in the marketplace. Expanding a sentiment analysis software is a way to utilise computers to measure consumer perceptions. Analyze the sentiment from employee tweets regarding work from home is the main goal. The employees work from home tweets dataset as input was collected from twitter. Then, to analyse the sentiment the NLP techniques, text classification and deep learning algorithm were used. The experimental results shows the performance metrics such as accuracy and analyse the sentiment based on sentiment analyser into positive, negative and neutral.KeywordSocial mediaSentiment analysisLong Short Term MemoryNLP techniquesTf-idf Vectorizer