Medium Gaussian SVM, Wide Neural Network and stepwise linear method in estimation of Lornoxicam pharmaceutical solubility in supercritical solvent

Prediction of methane hydrate equilibrium in saline water solutions based on support vector machine and decision tree techniques

The formation of clathrate hydrates offers a powerful approach for separating gaseous substances, desalinating seawater, and energy storage at low temperatures. On the other hand, this phenomenon may lead to practical challenges, including the blockage of pipelines, in some industries. Consequently, accurately predicting the equilibrium conditions for clathrate hydrate formation is crucial. This study was undertaken to design reliable models capable of predicting the equilibrium state of methane hydrates in saline water solutions. A comprehensive collection of measured data, consisting of 1051 samples, was assembled from published sources. The prepared databank encompassed the hydrate formation temperature of methane (HFTM) in the presence of 26 different saline water solutions. A machine learning modeling was undertaken through the implementation of Decision Tree (DT) and Support Vector Machine (SVM) approaches. While both models had excellent performance, the latter achieved higher accuracy in estimating the HFTM with the mean absolute percentage error (MAPE) of 0.26%, and standard deviation (SD) of 0.78% in the validation process. Furthermore, more than 90% of the values predicted by the novel models fell within the [Formula: see text]1% error bound. It was found that the intelligent models also favorably describe the physical variations of HFTM with operational factors. An examination using the William's plot acknowledged the truthfulness of the gathered data and the suggested estimation techniques. Ultimately, the order of significance of the factors governing the HFTM was clarified using a sensitivity analysis.

Atypical brain network development of infants at elevated likelihood for autism spectrum disorder during the first year of life

Autism spectrum disorder (ASD) is a neurodevelopmental disorder characterized by behavioral features that appear early in life. Although studies have shown that atypical brain functional and structural connectivity are associated with these behavioral traits, the occurrence and initial alterations of brain networks have not been fully investigated. The current study aimed to map early brain network efficiency and information transferring in infants at elevated likelihood (EL) compared to infants at typical likelihood (TL) for ASD in the first year of life. This study used a resting-state functional near-infrared spectroscopy (fNIRS) approach to obtain the length and strength of functional connections in the frontal and temporal areas in 45 5-month-old and 38 10-month-old infants. Modular organization and small-world properties were detected in both EL and TL infants at 5 and 10 months. In 5-month-old EL infants, local and nodal efficiency were significantly greater than age-matched TL infants, indicating overgrown local connections. Furthermore, we used a support vector machine (SVM) model to classify infants with or without EL based on the obtained global properties of the network, achieving an accuracy of 77.6%. These results suggest that infants with EL for ASD exhibit inefficiencies in the organization of brain networks during the first year of life.

Estimating the Dissolution of Anticancer Drugs in Supercritical Carbon Dioxide with a Stacked Machine Learning Model

Synthesizing micro-/nano-sized pharmaceutical compounds with an appropriate size distribution is a method often followed to enhance drug delivery and reduce side effects. Supercritical CO2 (carbon dioxide) is a well-known solvent utilized in the pharmaceutical synthesis process. Reliable knowledge of a drug's solubility in supercritical CO2 is necessary for feasible study, modeling, design, optimization, and control of such a process. Therefore, the current study constructs a stacked/ensemble model by combining three up-to-date machine learning tools (i.e., extra tree, gradient boosting, and random forest) to predict the solubility of twelve anticancer drugs in supercritical CO2. An experimental databank comprising 311 phase equilibrium samples was gathered from the literature and applied to design the proposed stacked model. This model estimates the solubility of anticancer drugs in supercritical CO2 as a function of solute and solvent properties and operating conditions. Several statistical indices, including average absolute relative deviation (AARD = 8.62%), mean absolute error (MAE = 2.86 × 10-6), relative absolute error (RAE = 2.42%), mean squared error (MSE = 1.26 × 10-10), and regression coefficient (R2 = 0.99809) were used to validate the performance of the constructed model. The statistical, sensitivity, and trend analyses confirmed that the suggested stacked model demonstrates excellent performance for correlating and predicting the solubility of anticancer drugs in supercritical CO2.