Classification of COVID19 Tweets based on Sentimental Analysis

The year 2020, has seen the advent of a pandemic that has affected the world as we know it globally. The origin reportedly from Wuhan, China, this pandemic is caused by COVID-19 which belongs to the family of Coronavirus. The increase of infection and mortality has shot up exponentially and has left mankind bewildered amongst the remains of the unseen disaster. During these times of hardship mankind has to face with a series of emotions. Analysis of all these emotions becomes a primary target for the well-being of an individual and mankind as a whole. The main motive of our study is to analyze these emotions correctly. Gathering these big chunks of data about this study from different social platforms like Twitter, Facebook, Instagram, etc. plays a major role. For this study we will be considering only the corona virus related tweets from Twitter. Analysis of all these tweets will give us a proper insight about the real emotions that the people has to face during these COVID-19 times. The main objective is to work with multinomial attributed to assess the sentiments more precisely. The next step is cleaning the data and labelling them for further processing. Hereafter a model is developed which is used to access the data and then predict the actual sentiment behind the tweet. The data is assessed using the binary-class and multi-class property with the cross-data evaluation of various machine learning algorithms to form the model. After tedious training of models, it is seen that the proposed model gives us a 96.58% accuracy with Support Vector Machine algorithm.