Sentimental Classification Method of Twitter Data for Indian Air Asia Services Analysis

EasyChair Preprint 3409

Abstract

In India there are many airline services which provide various types of services to their customers such as food beverage, important entertainment, Seat comfort, Staff Service and flight punctuality but they cannot express their feedback immediately so twitter provide a sound of data source for them to do customer sentiment analysis. In this research paper we focused on only Air Asia Airline services and collect data from twitter which is a huge SNS (Social Networking Site) with user posting by using WebHarvy tool. In our experiment, this analysis was carried out using 5 different classification strategies: Decision Tree, Random Forest, SVM, Logistic Regression, and Naive Bayes. The outcome of the test set is the tweet sentiment (positive/negative/neutral) with 3 class dataset and calculate the performance in terms of accuracy. We have achieved best accuracy 79.53% in case of Logistic Regression classifier. In this paper we are classifying sentiment of Twitter messages by exhibiting results of a machine learning algorithm using Rapid Miner. The tweets are extracted and pre-processed and then categorizing them in neutral, negative and positive sentiments finally summarizing the results as a whole.

Keyphrases: Machine Learning Techniques, Sentiment Analysis, Twitter Analysis, lemmatization

@booklet{EasyChair:3409,
  author    = {Rajat Yadu and Ragini Shukla},
  title     = {Sentimental Classification Method of Twitter Data for Indian Air Asia Services Analysis},
  howpublished = {EasyChair Preprint 3409},
  year      = {EasyChair, 2020}}