Abstract- Customer reviews are a powerful strategic tool for industry experts and for the Airline industry it deserves great attention. Before now, traditional methods used for this review provide inconsistent information about data in the airline industry. Sentiment analysis from online comments posted on review or micro-blogging sites has the potential to assist customers decide best airline of choice by analyzing other customer opinions. This paper presents information about airline reservation system using sentimental analysis with Naïve bayes classifier. The system uses airline information dataset, containing 16 columns of which few columns were selected by means of feature extraction. The selected columns are the text column with the airline sentiment column. Stop words were used in eliminating words that are so commonly used that they carry little information. LabelEncoder function was used in converting the y variable, which arranges the sentiment analysis columns into numbers ranging from -1, 0 to 1, where -1 represents a neutral comment, 1 represents positive and 0 represents negative. Further 70% of the dataset was used in training the Naïve bayes classifier while 30% of data was used for testing. Result show training accuracy of about 95% and a testing accuracy of about 79%. The Naïve bayes classifier was then saved and deployed to web using python flask.
Keywords- Airline Reservation, Sentiment Analysis, Naïve Bayes Classifier, Stopwords
|
DOI:
10.17148/IJARCCE.2021.10802
