Accurate forecasting of the Consumer Price Index (CPI) is significant for informed economic planning and decision-making. This project compares the performance of four machine and deep learning models in predicting Malaysia's CPI using various input structures.
- Support Vector Regression (SVR)
- Random Forest (RF)
- XGBoost
- Neural Networks (NN)
- Univariate Input: CPI only
- Multivariate Input: 12 economic indicators
- Simple Multivariate Input: Top 3 most correlated indicators
- SVR consistently outperformed other models across all input settings, particularly when using 12 multivariate indicators.
- In the multivariate setting, SVR achieved the best metrics:
- MAE: 0.0584
- RMSE: 0.0710
- R²: 0.99997
- Random Forest (RF) also showed strong performance in univariate and simplified multivariate configurations.
- Neural Networks (NN) demonstrated the weakest performance, particularly in multivariate settings (MAE up to 1.05).
SVR was applied to CPI datasets from Cambodia, Myanmar, and Laos to test cross-country performance.
It maintained:
- R² > 0.99
- MAE < 0.2
These results highlight SVR’s robustness across Southeast Asian economies.
- Python, Jupyter Notebook
- pandas, numpy, seaborn, matplotlib
- scikit-learn, XGBoost, Keras
- statsmodels, scikit-plot
FYP-CPI-Forecasting
- data/ # Datasets (Malaysia, Cambodia, Laos, Myanmar)
- notebooks/ # Model development and testing
- results/ # Visualizations and metrics
- README.md # Project documentation
This study demonstrates that machine learning—particularly Support Vector Regression (SVR) is highly effective in forecasting CPI, outperforming deep learning in both accuracy and consistency. The model's success across neighboring countries suggests its broader regional applicability.