Poultry Chicken Price Forecasting Using GRU–LSTM-Based Hybrid Deep Learning Model

Abstract

The precise poultry chicken price forecasting is necessary to maintain food security and market stability as well as protein demand mitigation in many countries. Existing machine learning-based forecasting models have problems with feature selection, hyperparameter tuning, linear assumptions, lower accuracy values, and insufficient data integration. Additionally, they did not offer highly accurate and month-based poultry chicken price forecasts. A deep learning (DL)-based framework for 30-day broiler chicken price forecasting is presented in this work. This work makes use of daily data from multiple sources, such as retail prices, feed costs, and macroeconomic variables. The proposed method combines ensemble feature selection, resilient feature engineering, and sophisticated data preparation, and it is optimized using walk-forward cross-validation and Bayesian optimization using Optuna. This work proposed a hybrid DL ensemble technique for monthly poultry price forecasting that combines Long Short-Term Memory (LSTM) networks with the Gated Recurrent Unit method. The proposed model outperforms previous studies by obtaining at least 2% lower MAPE and 10% higher R 2 value, with a MAPE score of 3.69% and R-squared value of 0.832. The feature contribution is explained in this work for forecasting using SHapley Additive exPlanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME) methods. 

Received: 8 August 2025 | Revised: 13 April 2026 | Accepted: 25 June 2026

Conflicts of Interest

The authors declare that they have no conflicts of interest to this work. 

Data Availability Statement

The data that support the findings of this study are openly available in Kaggle at https://www.kaggle.com/datasets/mhchowdhury37/poultry-chicken-dataset-new.

Author Contribution Statement

Mohammad Yasin Arafat: Methodology, Software, Validation, Formal analysis, Investigation, Resources. Mahfuzulhoq Chowdhury: Conceptualization, Data curation, Writing – original draft, Writing – review & editing, Visualization, Supervision, Project administration.