RESUMEN
Echocardiography, a simple and noninvasive tool, is the first choice for screening pulmonary hypertension (PH). However, accurate assessment of PH, incorporating both the pulmonary artery pressures and additional signs for PH remained unsatisfied. Thus, this study aimed to develop a machine learning (ML) model that can automatically evaluate the probability of PH. This cohort included data from 346 (275 for training set and internal validation set and 71 for external validation set) patients with suspected PH patients and receiving right heart catheterization. Echocardiographic images on parasternal short axis-papillary muscle level (PSAX-PML) view from all patients were collected, labeled, and preprocessed. Local features from each image were extracted and subsequently integrated to build a ML model. By adjusting the parameters of the model, the model with the best prediction effect is finally constructed. We used receiver-operating characteristic analysis to evaluate model performance and compared the ML model with the traditional methods. The accuracy of the ML model for diagnosis of PH was significantly higher than the traditional method (0.945 vs. 0.892, p = 0.027 [area under the curve [AUC]]). Similar findings were observed in subgroup analysis and validated in the external validation set (AUC = 0.950 [95% CI: 0.897-1.000]). In summary, ML methods could automatically extract features from traditional PSAX-PML view and automatically assess the probability of PH, which were found to outperform traditional echocardiographic assessments.
RESUMEN
Catalytic reactions are severely restricted by the strong adsorption of product molecules on the catalyst surface, where promoting desorption of the product and hindering its re-adsorption benefit the formation of free sites on the catalyst surface for continuous substrate conversion1,2. A solution to this issue is constructing a robust nanochannel for the rapid escape of products. We demonstrate here that MFI zeolite crystals with a short b-axis of 90-110 nm and a finely controllable microporous environment can effectively boost the Fischer-Tropsch synthesis to olefins by shipping the olefin molecules. The ferric carbide catalyst (Na-FeCx) physically mixed with a zeolite promoter exhibited a CO conversion of 82.5% with an olefin selectivity of 72.0% at the low temperature of 260 °C. By contrast, Na-FeCx alone without the zeolite promoter is poorly active under equivalent conditions, and shows the significantly improved olefin productivity achieved through the zeolite promoter. These results show that the well-designed zeolite, as a promising promoter, significantly boosts Fischer-Tropsch synthesis to olefins by accelerating escape of the product from the catalyst surface.
Asunto(s)
Zeolitas , Adsorción , Alquenos , Catálisis , Hierro , Zeolitas/químicaRESUMEN
Due to the serious economic losses and deaths caused by COVID-19, the modeling and control of such a pandemic has become a hot research topic. This paper finds an analogy between a polymerization reaction and COVID-19 transmission dynamics, which will provide a novel perspective to optimal control measures. Susceptible individuals, exposed people, infected cases, recovered population, and the dead can be assumed to be specific molecules in the polymerization system. In this paper, a hypothetical polymerization reactor is constructed to describe the transmission of an epidemic, and its kinetic parameters are regressed by the least-squares method. The intensity of social distancing u is considered to the mixing degree of the reaction system, and contact tracing and isolation ρ can be regarded as an external circulation in the main reactor to reduce the concentration of active species. Through these analogies, this model can predict the peak infection, deaths, and end time of the epidemic under different control measures to support the decision-making process. Without any measures (u = 1.0 and ρ = 0), more than 90% of the population would be infected. It takes several years to complete herd immunity by nonpharmacological intervention when the proportion of deaths is limited to less than 5%. However, vaccination can reduce the time to tens to hundreds of days, which is related to the maximum number of vaccines per day.
RESUMEN
Quantitative determination of polyvinyl chloride (PVC) concentration by Raman spectrum was studied in the present work. According to partial least squares (PLS) analysis, it was found that scores of PLS factor 1 were proportional to the concentrations of the sample solutions. Meanwhile, the loadings of factor 1 could reflect the contents of PVC and cyclohexanone simultaneously. The PLS regression model for PVC concentration prediction was built. The values of r and root mean square error (RMSE) between predictive results and actual values were 0.9963 and 2.775, respectively. The Raman characteristic peaks of PVC and cyclohexanone were found, including the C-Cl bond for PVC (620 and 695 cm(-1)) and the alicyclic ketone for cyclohexanone (1709 cm(-1)). By using internal standard method, another model for PVC concentration prediction was established, and the values of r and RMSE were 0.9941 and 3.151, respectively. The results indicated that it is feasible to use Raman spectrum to detect the PVC concentration, which is of significant importance to PVC recycling.
RESUMEN
Ethylene content in polypropylene was studied by Raman spectrum, combined with partial least squares (PLS) method. The comparison between Raman spectra for polyethylene and polypropylene was carried out, and the spectra between 50 and 600, 600 and 1600, and 2700 and 3100 cm(-1) were analyzed respectively. The models for ethylene content prediction were built, while the model based on 50-3600 cm(-1) spectra gave the best performance. The experiment indicated that Raman spectrum gave the similar predictive results as the near infrared (NIR) spectrum; the values of correlation coefficient (r), relative average deviation (RAD) and root mean square error (RMSE) between predictive results and actual values were 0.995, 2.65% and 0.319, respectively. According to PLS analysis, the loadings of factor 1 could reflect the relationship between the composition of polypropylene molecular chain and ethylene content, and ethylene content had a positive correlation with CH2 content, but a negative correlation with content of CH3, C-H, and C-C. The results indicated that it was feasible to detect the ethylene content in polypropylene by Raman spectrum.