This project focuses on predicting weather conditions using various machine learning models. It utilizes data from the Kelmarsh wind farm, including turbine data, status data, and ERA5 weather data. The primary objective is to develop a model that minimizes forecasting errors and enhances prediction accuracy. The Root Mean Squared Error (RMSE) is used as the primary evaluation metric.
The dataset consists of six different sources, including:
- Turbine data
- Status data
- Kelmarsh ERA5 weather data
- Weather-related parameters (e.g., temperature, wind speed, pressure)
- Time-series features (e.g., timestamps, historical trends)
- Operational data from turbines
The following machine learning models were used and compared based on their RMSE scores:
- CatBoost - Gradient boosting model optimized for categorical data.
- LightGBM - Efficient gradient boosting framework optimized for speed.
- Random Forest - Ensemble learning method using multiple decision trees.
- Decision Tree - Simple tree-based model for making predictions.
- K-Nearest Neighbors (KNN) - Distance-based model using nearest data points.
- Lasso Regression - Regularized regression technique reducing complexity.
- Linear Regression - Baseline model for trend estimation.
- Naive Forecaster - Simple forecasting model used for benchmarking.
The models were evaluated using RMSE, which is suitable for this problem because:
- It penalizes larger errors more than smaller ones.
- It provides an interpretable measure in the same unit as the target variable.
- It effectively captures forecasting accuracy.
MAE Scores (Scaled to Single Digits):
| Model | MAE Score |
|---|---|
| CatBoost | 7.22 |
| LightGBM | 7.57 |
| Random Forest | 8.15 |
| Decision Tree | 8.92 |
| KNN | 10.17 |
| Lasso | 10.30 |
| Linear | 8.79 |
| Naive Forecaster | 2.06 |
- Fine-tuning hyperparameters for better model performance.
- Exploring deep learning models (e.g., LSTMs) for time-series forecasting.
- Adding explainability using SHAP values for feature importance analysis.
- Nisshanth V G
- Department of Computational Intelligence,
- SRM Institute of Science and Technology, Tamil Nadu, India.
- Nevin Haniel R
- Department of Computational Intelligence,
- SRM Institute of Science and Technology, Tamil Nadu, India.
- Yuvan Nagalingam H
- Department of Computational Intelligence,
- SRM Institute of Science and Technology, Tamil Nadu, India.