In this research work, we have performed Machine Learning and Lexicon Based Techniques to identify and analyze user's expression or opinion on covid-19 vaccination from social media platform that is Twitter and acquainted the bulk tweets from 01 June 2021 to August 2021 using various twitter hashtags. Machine learning based Classifiers are used for investigating the evaluation performance of Algorithms. Real time datasets and machine Learning Algorithms are compared with Best Data classification Evaluation based on the size of train data also another approach is to investigating the polarity by using Lexicon Based approach for this Bing Liu Lexicons and Stanford University Lexicons are used. The global pandemic has created the medical emergency and stops the many regular activities. The whole world in the lockdown or quarantine to because Coronavirus disease. Among them, Covaxin, Covishield, Pfizer, Moderna and SputnikVare popular. Universally publicare articulating opinions on protection and success of the vaccines on social media. Research article shows, such tweets are collected from developer Application Management using a Twitter API. Unprocessed tweets are kept and preprocessed through Machine Learning techniques. Users opinion are predicted using a Classifiers Decision Tree, Support Vector Machine, K NN Algorithm and Naïve Bayes. Comparative machine learning classifiers study here comparative analysis is got highest accuracy of 97% for Decision tree with Covaxin dataset, Support vector machine with 94% for SputnikV, Naïve Bayes got highest accuracy of 95 for Covishield dataset and KNN got Highest accuracy of 96% for Covaxin. The Lexicon Based polarity classifies the score into three users opinions, positive, negative, and neutral. Result shows that, Covaxin shows 28.

In today's world an online existence and social media users utilize various social media platforms to express or comments their observations and opinions. The role of social media platforms are predicting Government Initiatives, Election results, product Analysis, business analysis, movie popularity, sports outcomes and stock market analysis. This review paper proposed the opinions are expressed through different social media platforms can be used for retrieving or extracting the real time predictions on several trends. As per the sentiment identification outcome find the features in the form of Positive (+ve), Negative (−ve) and Neutral (=). In this proposed research methodology, here collect user's reviews on particular trends, then preprocessed it, creation of the features and selecting for data classification using different machine leering classifiers and predict the result. For better performance, used advanced preprocessing techniques will be applied to cleaning the data. For Sentiment Classification will be used machine learning algorithms or techniques like (SVM) Support Vector machine, (ME) Maximum Entropy, (NB) Naïve Bayes and (DT) Decision tree. As per existing techniques, It is very difficult to mine the correct predictions from social media. Therefore, the prediction model will be designed for doing the prediction using real time data from Twitter. An opinion from text or comment posted on social media platforms by various categories of users is one of the critical and time consuming tasks in the field of opining mining and analysis. The importance of this proposed intelligent system for social media is to automatically providing polarity from unstructured data in the form of text in English language for effective decision making.