RETRACTED: N-Beats architecture for explainable forecasting of multi-dimensional poultry data.

Area of Science:

• Agricultural Science
• Data Science
• Artificial Intelligence

Background:

• Poultry production is vital to the agricultural economy, necessitating accurate data forecasting.
• Optimizing revenue, resource use, and productivity hinges on reliable poultry data predictions.

Purpose of the Study:

• To introduce a novel application of the N-BEATS architecture for multi-dimensional poultry data forecasting.
• To enhance forecast interpretability using an integrated Explainable AI (XAI) framework.
• To improve decision-making in poultry farm management through transparent and interpretable forecasts.

Main Methods:

• Applied the N-BEATS architecture for time series modeling.
• Utilized a multivariate dataset of environmental parameters for poultry disease diagnostics.
• Integrated an Explainable AI (XAI) framework for enhanced interpretability.

Main Results:

• N-BEATS outperformed conventional deep learning models (LSTM, GRU, RNN, CNN).
• Achieved low error metrics: MAE (0.172), RMSE (0.313), MSLE (0.042), RMSLE (0.204).
• Demonstrated robustness with a positive R-squared value (0.034), indicating superior performance.

Conclusions:

• N-BEATS is a superior and interpretable solution for complex, multi-dimensional forecasting in poultry production.
• The findings have significant implications for enhancing predictive analytics in agriculture.
• This approach offers a robust method for optimizing poultry farm management.