Rania M. Alhazmi
The pharmaceutical cold chain is a specialized unit in the logistics field that has demanding needs for transportation and warehousing in the supply chain to protect the reliability of pharmaceutical goods and guarantee the health security of individuals. The cost of the pharmaceutical cold chain is high. Recently, pharmaceutical drug traceability systems have been established as essential devices to develop the digital visibility and transparency of the supply chain. The complete pharmaceutical cold chain consists of a chain of links, such as storage and use, production, and transportation. Blockchain (BC)-based drug traceability offers an encouraging method for a distributed shared data platform that is reliable, permanent, and trustworthy. This paper proposes a Digital Transformation in Pharmaceutical Cold Chain Management using a Hybrid Deep Neural Network (DTPCCM-HDNN) model, intending to enhance digital transformation and transparency in pharmaceutical cold chain management by utilizing BC and advanced models. Initially, the data preprocessing stage employs the RobustScaler method to clean, transform, and structure raw data into an appropriate format. The feature selection process involves the Mutual Information (MI), Recursive Feature Elimination (RFE), and Random Forest Importance (RFI) methods to choose crucial features from the dataset. Finally, a hybrid Convolutional Neural Network and Long Short-Term Memory (CNN-LSTM) technique is implemented for prediction. The performance of the DTPCCM-HDNN approach was examined on a pharmaceutical SC optimization dataset. The DTPCCM-HDNN approach demonstrated great performance, achieving an MSE of 0.0673, an RMSE of 0.2595, and an MAE of 0.2245.