Dataset and Code for SKU-Level Demand Forecasting Using ML–PPCP Framework

Description

Synthetic dataset and Python code provided to ensure reproducibility of the machine learning and deterministic production planning framework proposed in the associated manuscript. Real industrial data are confidential; therefore, anonymized synthetic data are supplied.

Steps to reproduce

Reproducibility steps The following steps allow full replication of the proposed forecasting workflow: Download the dataset and source code from the public repository. Install the required dependencies listed in requirements.txt. Execute the training script: python mlp_training_example.py The script automatically: loads the synthetic dataset, splits training and testing sets, trains a Multilayer Perceptron regressor, evaluates the model using MAE, RMSE, and R² metrics. The experiment can be executed either locally or in cloud environments such as Google Colab. All results are deterministic given the fixed random seed.

Institutions

• Universidade Federal de Santa Catarina

  SC, Florianopolis

Categories

Licence