Flight Price Prediction Using Machine Learning

For some rationale, anyone who's thought about lunchtime getting gone lighter in the execute must have observed by now that price prediction lies truly at the core of the airline industry. Hence, this work carries out behavior analysis of the DT, RF, and Logistic Regression algorithms on price prediction for flights. For starters, Decision Trees are considered to be simple and easy to explain by laymen; hence, they are studied first. Such trees learn price patterns but can merely approximate relationships when such relationships are highly complicated in the dataset. Then Random Forest being an ensemble method ought to be studied for the possibility of improving the prediction by combining many decision trees. Finally, Logistic Regression is adapted for flight price prediction since it is commonly used as a classification method and gains from doing so when tackling the binary outcome feature of tourist tickets. Flight pricing data is real-world, and the performance of each algorithm is checked in terms of r2 score. Furthermore, a feature importance analysis is done to identify aspects affecting flight prices. Hence, through these three machine-learning methods, this study aims to impart an idea of which machine-learning method affords strengths and weaknesses in the opponent of flight price prediction. This finding will help the stakeholders, from the travelers to the airlines, make a better decision.