Prediction of Airline Ticket Price Using Machine Learning Method
Hüseyin KorkmazAirline ticket pricing is a complex and dynamic process influenced by various factors, including demand fluctuations, seasonal variations, and competitive strategies. Accurate price prediction is crucial for both airlines, to maximize revenue, and customers, to secure the best deals. Traditional methods often fall short of capturing the intricate and rapidly changing patterns of airfare pricing. With the advent of machine learning algorithms, there is a growing potential to enhance the accuracy and reliability of ticket price predictions. This paper aims to predict ticket prices based on airline flight data using ML algorithms and to compare the performance of ML algorithms. The secondary objective of this paper is to identify the main factors affecting airline ticket prices. The flight and ticket price datasets of THY and PGS that were obtained from open-access sources are used in this paper. The final dataset consists of 962 records for three months from June 1st, 2022 to August 30th, 2022 and includes 19 different variables. Statistical tests and ML algorithms were applied to the final dataset. This paper compares various ML models to predict airline ticket prices, considering performance metrics such as MAE, MSE, RMSE, and R2 during training and test phases. According to the model training and test results, the best algorithm is GPR with R2: 0.86 (training) and R2: 0.90 (test). The findings are consistent with existing literature, further validating the superior efficacy of certain models in specific contexts and demonstrating significant progress in the field. This paper contributes to the literature by comparing the effectiveness of various machine learning algorithms in predicting airline ticket prices, providing new and valuable insights into model performance and key price-determining factors.
