Search Results (6)

The lack of clinical data for chronic kidney disease (CKD) prediction frequently results in model overfitting and inadequate generalization to novel samples. This research mitigates this constraint by utilizing a Conditional Tabular Generative Adversarial Network (CTGAN) to enhance a constrained CKD dataset sourced from the University of California, Irvine (UCI) Machine Learning Repository. The CTGAN model was trained to produce realistic synthetic samples that preserve the statistical and feature distributions of the original dataset. Multiple machine learning models, such as AdaBoost, Random Forest, Gradient Boosting, and K-Nearest Neighbors (KNN), were assessed on both the original and enhanced datasets with incrementally increasing degrees of synthetic data dilution. AdaBoost attained 100% accuracy on the original dataset, signifying considerable overfitting; however, the model exhibited enhanced generalization and stability with the CTGAN-augmented data. The occurrence of 100% test accuracy in several models should not be interpreted as realistic clinical performance. Instead, it reflects the limited size, clean structure, and highly separable feature distributions of the UCI CKD dataset. Similar behavior has been reported in multiple previous studies using this dataset. Such perfect accuracy is a strong indication of overfitting and limited generalizability, rather than feature or label leakage. This observation directly motivates the need for controlled data augmentation to introduce variability and improve model robustness. The dataset with the greatest dilution, comprising 2000 synthetic cases, attained a test accuracy of 95.27% utilizing a stochastic gradient boosting approach. Ensemble learning techniques, particularly gradient boosting and random forest, regularly surpassed conventional models like KNN in terms of predicted accuracy and resilience. The results demonstrate that CTGAN-based data augmentation introduces critical variability, diminishes model bias, and serves as an effective regularization technique. This method provides a viable alternative for reducing overfitting and improving predictive modeling accuracy in data-deficient medical fields, such as chronic kidney disease diagnosis. Full article

Recent applications of low-field NMR in alkali-activated materials (AAMs) often adopt interpretation models developed for Portland cement systems, overlooking the distinct influences of paramagnetic/ferrimagnetic components and free-water redistribution. This study investigates how paramagnetic or ferrimagnetic component and free water distribution influence low-field nuclear magnetic resonance (LF-NMR) and proton density magnetic resonance imaging (PD-MRI) characterization of alkali-activated materials (AAMs). Blast furnace slag, fly ash, and steel slag were activated with NaOH solution at liquid-to-solid ratios of 0.45 and 0.5, and analyzed across top, middle, and bottom layers. Slurries prepared with less mixing water and CaO-rich raw materials exhibited negligible settling and uniform relaxation behavior, whereas those with higher water content and CaO-deficient raw materials showed pronounced stratification, resulting in distinct gradients in signal intensity. The results indicate that the LF-NMR data interpretation of relatively dilute system may be unreliable as the relaxation time of protons will be extended after they transfer from bottom to the top of the slurry. A preliminary method for assessing slurry suitability for LF-NMR characterization is proposed for future validation. Full article

Background: Fosfomycin (FOS) is an older antimicrobial agent newly rediscovered as a possible treatment for infections with limited therapeutic options (e.g., Gram-negative bacteria with difficult-to-treat resistance, DTR), especially in intravenous form. However, for correct usage of FOS, it is necessary to have a reliable susceptibility testing method suitable for routine practice and robust interpretation criteria. Results: The results were interpreted according to 2023 interpretation criteria provided by the European Committee on Antimicrobial Susceptibility Testing (EUCAST). DTR Gram-negatives were more likely to be resistant to FOS (45% in Enterobacterales and 20% in P. aeruginosa) than non-DTR (10% and 6.7%, resp.). All isolates of S. aureus were susceptible to FOS. In Gram-negatives, all agreement values were unacceptable. Etest^® performed better in the DTR cohort (categorical agreement, CA, 80%) than in the non-DTR cohort (CA 45.7%). There were no very major errors (VREs) observed in P. aeruginosa. S. aureus had surprisingly low essential agreement (EA) rates (53% for MRSA and 47% for MSSA) for Etest^®, but categorical agreement was 100%. Methods: A total of 130 bacterial isolates were tested and compared using the disc diffusion method (DD) and gradient strip method (Etest^®) with the reference method (agar dilution, AD). The spectrum of isolates tested was as follows: 40 Enterobacterales (20 DTR vs. 20 non-DTR), 30 Pseudomonas aeruginosa (15 DTR vs. 15 non-DTR), and 60 Staphylococcus aureus (30 methicillin-susceptible, MSSA, vs. 30 methicillin-resistant, MRSA). Conclusions: Neither one of the tested methods was identified as a suitable alternative to AD. It would be beneficial to define more interpretation criteria, at least in some instances. Full article

The past decade has witnessed the rapid development of machine learning applied in economics and finance. Recent evidence suggests that machine learning models have produced superior results to traditional statistical models and have become the driving force for dramatic improvement in the financial industry. However, a much-debated question is whether the prediction results from black box machine learning models can be interpreted. In this study, we compared the predictive power of machine learning algorithms and applied SHAP values to interpret the prediction results on the dataset of listed companies in Vietnam from 2010 to 2021. The results showed that the extreme gradient boosting and random forest models outperformed other models. In addition, based on Shapley values, we also found that long-term debts to equity, enterprise value to revenues, account payable to equity, and diluted EPS had greatly influenced the outputs. In terms of practical contributions, the study helps credit rating companies have a new method for predicting the possibility of default of bond issuers in the market. The study also provides an early warning tool for policymakers about the risks of public companies in order to develop measures to protect retail investors against the risk of bond default. Full article

Inefficiency of medical therapies used in order to cure patients with bacterial infections requires not only to actively look for new therapeutic strategies but also to carefully select antibiotics based on variety of parameters, including microbiological. Minimal inhibitory concentration (MIC) defines in vitro levels of susceptibility or resistance of specific bacterial strains to applied antibiotic. Reliable assessment of MIC has a significant impact on the choice of a therapeutic strategy, which affects efficiency of an infection therapy. In order to obtain credible MIC, many elements must be considered, such as proper method choice, adherence to labeling rules, and competent interpretation of the results. In this paper, two methods have been discussed: dilution and gradient used for MIC estimation. Factors which affect MIC results along with the interpretation guidelines have been described. Furthermore, opportunities to utilize MIC in clinical practice, with pharmacokinetic /pharmacodynamic parameters taken into consideration, have been investigated. Due to problems related to PK determination in individual patients, statistical estimation of the possibility of achievement of the PK/PD index, based on the Monte Carlo, was discussed. In order to provide comprehensive insights, the possible limitations of MIC, which scientists are aware of, have been outlined. Full article

The in vitro interactions of isavuconazole in combination with colistin were evaluated against 55 clinical Aspergillus species isolates belonging to the five most important species (Aspergillus flavus, Aspergillus fumigatus, Aspergillus nidulans, Aspergillus niger, and Aspergillus terreus) responsible for human aspergillosis by a microdilution checkerboard technique based on the European Committee on Antimicrobial Susceptibility Testing (EUCAST) reference method for antifungal susceptibility testing. Selected isolates (A. nidulans, n = 10; A. niger, n = 15) were additionally evaluated by an agar diffusion assay using isavuconazole gradient concentration strips with or without colistin incorporated Roswell Parc Memorial Institute (RPMI) agar. Interpretation of the checkerboard results was done by the fractional inhibitory concentration index. Using the checkerboard method, combination isavuconazole–colistin was synergistic for 100% of the 15 A. nidulans isolates and for 60% of the 20 A. niger isolates. No interactions were found for any of the other isolates. By agar diffusion assay, minimal inhibitory concentrations (MICs) in combination decreased compared to isavuconazole alone for 92% of the isolates. No interactions were found for any A. nidulans isolates, but synergy was observed for 40% of the A. niger isolates. A poor essential agreement of EUCAST and gradient concentration strip MICs at ± 2 log₂ dilutions with 0% was obtained. Antagonistic interactions were never observed regardless of the technique used. Full article