Development of GBRT Model as a Novel and Robust Mathematical Model to Predict and Optimize the Solubility of Decitabine as an Anti-Cancer Drug

Modeling and validation of drug release kinetics using hybrid method for prediction of drug efficiency and novel formulations

This paper presents a thorough examination for drug release from a polymeric matrix to improve understanding of drug release behavior for tissue regeneration. A comprehensive model was developed utilizing mass transfer and machine learning (ML). In the machine learning section, three distinct regression models, namely, Decision Tree Regression (DTR), Passive Aggressive Regression (PAR), and Quadratic Polynomial Regression (QPR) applied to a comprehensive dataset of drug release. The dataset includes r(m) and z(m) inputs, with corresponding concentration of solute in the matrix (C) as response. The primary objective is to assess and compare the predictive performance of these models in finding the correlation between input parameters and chemical concentrations. The hyper-parameter optimization process is executed using Sequential Model-Based Optimization (SMBO), ensuring the robustness of the models in handling the complexity of the controlled drug release. The Decision Tree Regression model exhibits outstanding predictive accuracy, with an R2 score of 0.99887, RMSE of 9.0092E-06, MAE of 3.51486E-06, and a Max Error of 6.87000E-05. This exceptional performance underscores the model's capability to discern intricate patterns within the drug release dataset. The Passive Aggressive Regression model, while displaying a slightly lower R2 score of 0.94652, demonstrates commendable predictive capabilities with an RMSE of 6.0438E-05, MAE of 4.82782E-05, and a Max Error of 2.36600E-04. The model's effectiveness in capturing non-linear relationships within the dataset is evident. The Quadratic Polynomial Regression model, designed to accommodate quadratic relationships, yields a noteworthy R2 score of 0.95382, along with an RMSE of 5.6655E-05, MAE of 4.49198E-05, and a Max Error of 1.86375E-04. These results affirm the model's proficiency in capturing the inherent complexities of the drug release system.

Simulation and Optimization: A New Direction in Supercritical Technology Based Nanomedicine

In recent years, nanomedicines prepared using supercritical technology have garnered widespread research attention due to their inherent attributes, including structural stability, high bioavailability, and commendable safety profiles. The preparation of these nanomedicines relies upon drug solubility and mixing efficiency within supercritical fluids (SCFs). Solubility is closely intertwined with operational parameters such as temperature and pressure while mixing efficiency is influenced not only by operational conditions but also by the shape and dimensions of the nozzle. Due to the special conditions of supercriticality, these parameters are difficult to measure directly, thus presenting significant challenges for the preparation and optimization of nanomedicines. Mathematical models can, to a certain extent, prognosticate solubility, while simulation models can visualize mixing efficiency during experimental procedures, offering novel avenues for advancing supercritical nanomedicines. Consequently, within the framework of this endeavor, we embark on an extensive review encompassing the application of mathematical models, artificial intelligence (AI) methodologies, and computational fluid dynamics (CFD) techniques within the medical domain of supercritical technology. We undertake the synthesis and discourse of methodologies for calculating drug solubility in SCFs, as well as the influence of operational conditions and experimental apparatus upon the outcomes of nanomedicine preparation using supercritical technology. Through this comprehensive review, we elucidate the implementation procedures and commonly employed models of diverse methodologies, juxtaposing the merits and demerits of these models. Furthermore, we assert the dependability of employing models to compute drug solubility in SCFs and simulate the experimental processes, with the capability to serve as valuable tools for aiding and optimizing experiments, as well as providing guidance in the selection of appropriate operational conditions. This, in turn, fosters innovative avenues for the development of supercritical pharmaceuticals.

Computational intelligence modeling of hyoscine drug solubility and solvent density in supercritical processing: gradient boosting, extra trees, and random forest models

This work presents the results of using tree-based models, including Gradient Boosting, Extra Trees, and Random Forest, to model the solubility of hyoscine drug and solvent density based on pressure and temperature as inputs. The models were trained on a dataset of hyoscine drug with known solubility and density values, optimized with WCA algorithm, and their accuracy was evaluated using R², MSE, MAPE, and Max Error metrics. The results showed that Gradient Boosting and Extra Trees models had high accuracy, with R² values above 0.96 and low MAPE and Max Error values for both solubility and density output. The Random Forest model was less accurate than the other two models. These findings demonstrate the effectiveness of tree-based models for predicting the solubility and density of chemical compounds and have potential applications in determination of drug solubility prior to process design by correlation of solubility and density to input parameters including pressure and temperature.