Sentimental Analysis of airline tweets using Machine Learning Algorithm and Regular Expression

Sentiment analysis is a technique that uses natural language processing techniques to analyze the feelings and usage of textual content evaluation.Sentiment analysis makes use of information identified and taken from a variety of sources. It looks at a variety of factors to try and ascertain a customer's attitude. It is extensively used throughout the world for sentiment analysis and opinion extraction, enabling business organizations to understand user preferences. This paper presents a thorough yet user-friendly method for evaluating sentiment, which primarily displays aggregate ratings based on both positive and negative words. Based on the ratings, the tweets are then classified as Positive, Negative, or Neutral. Over the past ten years, there has been a notable increase in the application of opinion mining and sentiment evaluation techniques. Many studies in this field seek to determine people's attitudes, beliefs, and feelings about a particular topic or individual. The evaluation's primary objective is to use a machine learning model to ascertain the sentiment of the reviews, and then to compare the outcome with a manual review of the data. This could help scientists represent and analyze assessments across domains and impartially support data evaluation. For review evaluation, a hybrid approach combining a supervised machine learning algorithm with natural language processing methods is suggested. Finding the first-class version is the aim of this mission in order to gauge the sentiment of airline tweets. Throughout the process of conducting research and considering various approaches and factors to consider, we noticed that approaches such as naive bayes and logistic regression have not been thoroughly investigated since along with data imbalancing was also not look into consideration.So We will assess a more practical method in this mission that makes use of regular expression, oversampling and ml models and look into data balancing.