Skip to content

henryph24/rmit_tsresearch

Repository files navigation

Rice Price Forecasting Project

Overview

This project implements multiple forecasting approaches to predict Vietnamese rice prices using both statistical and machine learning models. It combines macro-economic indicators, news sentiment data, and historical price data to generate forecasts.

Data Sources

  • target_rice_data.xlsx: Vietnamese 5% broken rice prices (monthly)
  • macro_data.xlsx: Macro-economic indicators from Vietnam
  • news_data_loaded.xlsx: News sentiment data
  • Time period: 2009-2024

Models Implemented

  1. Statistical Models (statsforecast.ipynb):

    • AutoARIMA
    • AutoETS
    • AutoTheta
    • AutoCES
  2. Machine Learning Models (mlforecast.ipynb):

    • ElasticNet
    • XGBoost
    • LightGBM
    • CatBoost

Key Features

  • Data preprocessing and cleaning
  • Automatic TS feature engineering & LLM-powered sentiment analysis
  • Cross-validation with rolling windows for most robustness checks
  • Ray-powered parallel processing for efficient statistical model training
  • Efficient automated ML & DL forecasting pipeline
  • Multiple evaluation metrics:
    • RMSE (Root Mean Square Error)
    • Directional Accuracy
    • Turning Point Accuracy
    • Weighted Combined Score

Project Structure

preprocessing.ipynb

  • Data loading and cleaning
  • Feature engineering
  • Data merging and preparation

mlforecast.ipynb

  • Machine learning model implementation
  • Model training and evaluation
  • Cross-validation
  • Results visualization

statsforecast.ipynb

  • Statistical model implementation
  • Model training and evaluation
  • Cross-validation
  • Results comparison

Requirements

Python 3.10+ Key packages:

  • pandas
  • numpy
  • scikit-learn
  • statsforecast
  • mlforecast
  • ray
  • matplotlib
  • xgboost
  • lightgbm
  • catboost

Usage

  1. Run preprocessing.ipynb first to prepare the data
  2. Run either statsforecast.ipynb or mlforecast.ipynb for predictions
  3. Results include performance metrics and visualizations

Note: The project uses parallel processing through Ray for efficient computation of statistical models.

License

This project is for research purposes only. Please ensure you have the necessary rights to use the data sources mentioned above.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published