AIRLINE PASSENGER SATISFACTION PREDICTION USING MACHINE LEARNING ALGORITHMS

Abstract

Airlines are always focused on providing excellent customer service. It is vital to determine which of the two levels of satisfaction with the airline the passenger belongs to: satisfied or dissatisfied. Customer satisfaction scores from 120,000+ airline passengers, including additional information about each passenger, have been collected to determine the factors that are highly correlated to a satisfied or dissatisfied passenger to predict the level of satisfaction the passenger belongs. This study presents a machine learning approach to analyze the information and improve the customer’s experience. The dataset has 24 columns of data, some are categorical, some are integer, one is float and many contain discrete values between 0-5. This paper contains various methods and techniques that we have used to determine the best accurate model and various visualization tools to deeply understand the data. This paper also helps to understand which features are important in predicting passenger satisfaction and to find the correlation between features. We have implemented various machine learning classification algorithms of which we observed that the Random Forest Classifier is the best model giving us the best accuracy result of 94% further helping in accurate predictions among all other models we have used. Hence, this could help airline companies adjust their service value and demand to satisfy customers’ demands.