<![CDATA[This study aims to develop a dynamic prediction model for airline ticket prices using the Multiple Linear Regression algorithm. The research utilizes the public dataset Flight Price Prediction from Kaggle, which originally contained 116,464 rows and 12 columns. After data cleaning by removing missing values (dropna()) and non-predictive columns (such as Flight_ID), the final dataset used for analysis consisted of 116,463 rows and 10 columns. Data preprocessing included handling missing data, encoding categorical variables, feature engineering, standardization, and multicollinearity testing using the Variance Inflation Factor (VIF). The MLR model achieved an R² of 0.882, MAE of 4573.37, and RMSE of 7797.53, indicating strong predictive performance for a linear model. The most influential factors were airline type, service class, number of stops, duration, and booking lead time. Full-service airlines such as Vistara and Air India tend to have higher ticket prices, while early bookings and economy class tickets significantly lower prices. The findings confirm that MLR remains a reliable baseline for interpretable, efficient, and explainable price forecasting systems. Future research may combine MLR with non-linear algorithms (e.g., Random Forest or Neural Network) to enhance accuracy. This study contributes to integrating data science into predictive information systems for dynamic airline pricing and decision support optimization.]]>
Copyrights © 2025
