No abstract

No abstract

Aqueous hydrolysis is used to chemically recycle polyethylene terephthalate (PET) due to production of high‐quality terephthalic acid (TPA), the PET monomer. PET hydrolysis depends on various factors including PET size, catalyst concentration, and reaction temperature. So, modeling PET hydrolysis by considering the effective factors can provide useful information for material researchers to specify how to design and run these reactions. It will save time, energy, and materials by optimizing the hydrolysis conditions. Machine learning algorithms enable to design models to predict the output results. For the first time, 381 experimental data were gathered to model aqueous hydrolysis of PET. Effective factors on PET hydrolysis were connected to the TPA yield. The logistic regression was applied to rank the effective factors. Two algorithms were proposed, artificial neural network multi‐layer perceptron (ANN‐MLP) and adaptive network‐based fuzzy inference system (ANFIS). The dataset was divided into training, validating, and testing sets to train, validate, and test the models, respectively. The models predicted TPA yield sufficiently where the ANFIS model outperformed. R‐squared (R2) and Root Mean Square Error (RMSE) loss functions were employed to measure the efficiency of the models and evaluate their performance.

Introduction : Previous studies have suggested that 24‐hour NIHSS may serve as a surrogate marker for functional outcomes in acute ischemic stroke patients. Here, we examine if 24‐hour NIHSS is a predictor of 90‐day mRS in the prospective Systematic Evaluation of Patients Treated With Neurothrombectomy Devices for Acute Ischemic Stroke (STRATIS) Registry. Methods : Data from the STRATIS Registry, a multicenter, non‐randomized, observational study of AIS LVO patients treated with the Solitaire stent‐retriever as the first‐choice therapy within 8 hours from symptoms onset, were analyzed. Patients with premorbid mRS>2, posterior circulation stroke, missing 24 NIHSS or 90‐day mRS were excluded from the analysis. The ability of 24‐hour NIHSS (continuous and thresholds ≤6 and ≤8) to predict 90‐day mRS using logistic regression was examined. The models were adjusted for age, baseline NIHSS, hypertension, diabetes, atrial fibrillation, IV‐tPA use, time to recanalization, and revascularization status. Results : Of the 938 STRATIS patients, 662 with 24‐hour NIHSS and 90‐day mRS were included. A model trained with the continuous 24‐hours NIHSS had higher predictive power (sensitivity 0.89, specificity 0.76, AUC 0.89±0.013, P<0.001) than the models trained with thresholds ≤6 and ≤8. Conclusions : When adjusted for covariates, 24‐hour NIHSS as a continuous variable was the strongest predictor of dichotomous mRS outcomes in STRATIS patients. Twenty‐four hour NIHSS ≤6 and ≤8 present the second and the third best results, respectively.

Class imbalance refers to a major issue in data mining where data with unequal class distribution can deteriorate classification performance. Although it alone affects the performance of the classifiers, the joint‐effect of class imbalance and overlap is more damaging. Data overlap happens when multiple classes are assigned to a single data point causing the classifiers to misidentify the class boundaries. This study offers a deep insight into the intricacies of the UNSW‐NB15 dataset and two issues that may lead the data‐driven models to demonstrate poor performance. The most commonly used visualization methods such as bar chart, 3D and 2D scatter plots, intercluster distance map, and parallel coordinate diagram were employed to depict the data imbalanced and overlap. However, their limitations in capturing the overlapping issue led us to propose an accurate, easy‐to‐interpret, and scalable overlapping visualization method. The method clearly detects the data overlap and illustrates the effect of several data scalers in dealing with the data overlap. To verify the accuracy of the proposed method, a number of classifiers were implemented along with the scalers and the calculated AUC scores were compared to those calculated from the classifiers that were implemented on the original dataset.