Sentiment Analysis of Twitter Data for Better Services of Airlines Using Classifier-Model Algorithms, AI, and NLP Techniques

Abstract

The rising attractiveness of social media sites and usage of social websites is to share information. Tweet, for instance, is a stage in which audiences direct, and examine posts recognized as “tweets,” and engage together in one-of-a-kind communities through feedback, comments, tweets, and reviews. There are so many companies that want tweets, comments, reviews, and analyses on their products and services, so they can improve their services, products, and customer satisfaction. We are highlighting airline companies where people are interested in checking the first company’s feedback, reviews, and tweets so they can use airline services as per their choice and needs with satisfaction. There are many airlines that have some good products and services, and some are bad, some are good for timings, luggage, meal, and customer service, Users proportion their everyday lives, and put up their reviews on all products and services. Many corporations can advantage profit from the large-scale platform by using amassing and analyzing data related to reviews on them. The foremost objective of this study article is to provide a classifier model; it may carry out sentimental and intent analysis of actual records accumulated from Tweets for airline services. Unstructured data in Tweet is the highly unstructured text which builds it harder to identify. However, in this article, we are classifying the sentiment and intent of tweets by revealing results through the classifier algorithms using python and its libraries, framework, and tools. We proposed our models based on NLP (natural language processing). Then, tweets are mined and pre-processed and then categorized into neutral, negative, and positive and highlight labels of services of the intent analysis it is a problem-solving analysis that helps people who want to use airline services, it would provide better recommendations as per people’s need. Naive Bayes classifier, Support Vector Machine, Random Forests, Logistic Regression, and Ensemble learning Voting Classifier algorithms are used by us for developing and outperform classifiers with better accuracy.