Predicting Airline Fare Prices using Random Forest: A Modular Approach for Cost-Effective Travel Planning

This study presents a machine learning-based approach for predicting airline fare prices using the random forest (RF) algorithm. The primary objective is to assist travelers in making cost-effective booking decisions by forecasting flight ticket prices based on key factors such as departure location, journey duration, airline, time of travel, and other relevant features. The dataset used for this study consists of 10,683 records, encompassing both local and international flights within India in 2019. The data underwent rigorous preprocessing, including handling missing values, encoding categorical variables, and performing feature engineering. Exploratory data analysis provided valuable insights into the relationships between different features and their influence on ticket prices. To enhance model performance, hyperparameter tuning was conducted using Randomized Search Cross-Validation (RandomizedSearchCV), resulting in an accuracy improvement from 79% to 81%. The model was validated using various metrics, including mean absolute error (MAE), root mean squared error (RMSE), and R² score, demonstrating its robustness and predictive capability. The results highlight the potential of machine learning, particularly the Random Forest algorithm, in predicting flight fare prices and assisting travelers in planning cost-effective journeys. Future work will explore incorporating external factors like real-time weather conditions and economic indicators, as well as expanding the model to include international flight data and integrating with booking platforms for a comprehensive travel planning solution.