Quantitative analysis of ferrohydrodynamics of blood containing magnetic nanocarriers for advanced drug delivery design via hybrid machine learning

Ferrohydrodynamics of blood containing magnetic nanocarriers was carried out in this study to evaluate the influence of magnetic field on the nanocarrier drug delivery formulation, with the goal of improving spatial control and efficiency in advanced targeted drug delivery systems. The model of magnetic field was carried out by Maxwell' model, while the blood velocity through the vessel was modeled using Navier-Stokes equations. Then the results of computations were used as input data for development of machine learning (ML) models. Indeed, a comprehensive study on the prediction of velocity (U) based on input variables (x, y) using three different ML models: K-Nearest Neighbor (KNN), Decision Tree (DT), and Gradient Boosting (GB) was studied. To achieve this, a systematic approach is followed, including pre-processing tasks such as normalization, outlier detection, and data splitting into training and testing sets. Hyper-parameter optimization is performed using the Rain Optimization Algorithm (ROA) to fine-tune the models and improve their predictive capabilities. The final findings show the robustness of the DT, KNN, and GB models in predicting velocity. The DT model achieved an R² value of 0.90278, highlighting a robust correspondence between model and expected values. The KNN model outperformed with an R² of 0.99088, showcasing its superior predictive accuracy, which is critical for precisely predicting nanoparticle trajectories and optimizing magnetic guidance in targeted cancer drug delivery. The GB model also exhibited strong performance with an R² score of 0.96168.