Comparison Of Models Research Articles

Life-course perspective on the causal mechanism of social functioning in schizophrenia spectrum disorder.

Due to prior emphasis on clinical recovery in Schizophrenia Spectrum Disorder (SSD), improving social functioning (SF) was oftentimes neglected, with ⩽15% of patients achieving social recovery. Priorly, we and others have shown that life-course factors, including childhood adversities, play a role in the occurrence and severity of postmorbid SF impairments, highlighting the need to understand these factors for effective interventions. This study investigates the mechanisms influencing SF in SSD and examines the causal roles of childhood trauma, premorbid adjustment, perceived stigma, self-esteem, and quality of life. This longitudinal study utilized data of 1,057 SSD patients, with measures at baseline, 3 and 6-year follow-up, from the Genetic Risk and Outcome of Psychosis cohort, to unravel the causal mechanism underlying SF utilizing Structural Equation Modeling. Determinants were assessed using validated retrospective and self-report questionnaires. Model development and testing involved a multistage process, encompassing relationships exploration, fit evaluation and model comparison. We developed a probable causal model. Impaired premorbid adjustment emerged as a crucial factor, exerting negative influence on long-term SF, with a direct effect of β = -.252, p < .001 on SF at 3-year follow-up and β = -.056, p = .073 at 6-year follow-up. Childhood trauma exhibited a negative direct effect on SF at 3-year follow-up (β = -.087, p = .039), while demonstrating a substantial carry-over, indirect effect. Elevated perceived stigma negatively affected SF at 3-year follow-up (β = -.112, p = .008). The model explained 9.9% of SF variation at the 3-year follow-up and 54.3% at the 6-year follow-up. While validation is necessary, we found a foundational basis for causal interpretation. Premorbid adjustment, childhood trauma, and perceived stigma are essential life-course factors shaping postmorbid SF in SSD. We recommend interventions covering both prevention and treatment for individuals with established symptoms-addressing childhood trauma, premorbid adjustment, and perceived stigma-to enhance long-term social outcomes and offer actionable insights for clinicians.

Biased expectations about future choice options predict sequential economic decisions

Considerable research has shown that people make biased decisions in “optimal stopping problems”, where options are encountered sequentially, and there is no opportunity to recall rejected options or to know upcoming options in advance (e.g. when flat hunting or choosing a spouse). Here, we used computational modelling to identify the mechanisms that best explain decision bias in the context of an especially realistic version of this problem: the full-information problem. We eliminated a number of factors as potential instigators of bias. Then, we examined sequence length and payoff scheme: two manipulations where an optimality model recommends adjusting the sampling rate. Here, participants were more reluctant to increase their sampling rates when it was optimal to do so, leading to increased undersampling bias. Our comparison of several computational models of bias demonstrates that many participants maintain these relatively low sampling rates because of suboptimally pessimistic expectations about the quality of future options (i.e. a mis-specified prior distribution). These results support a new theory about how humans solve full information problems. Understanding the causes of decision error could enhance how we conduct real world sequential searches for options, for example how online shopping or dating applications present options to users.

Comparison of Vendor-Pretrained and Custom-Trained Deep Learning Segmentation Models for Head-and-Neck, Breast, and Prostate Cancers

Background/Objectives: We assessed the influence of local patients and clinical characteristics on the performance of commercial deep learning (DL) segmentation models for head-and-neck (HN), breast, and prostate cancers. Methods: Clinical computed tomography (CT) scans and clinically approved contours of 210 patients (53 HN, 49 left breast, 55 right breast, and 53 prostate cancer) were used to train and validate segmentation models integrated within a vendor-supplied DL training toolkit and to assess the performance of both vendor-pretrained and custom-trained models. Four custom models (HN, left breast, right breast, and prostate) were trained and validated with 30 (training)/5 (validation) HN, 34/5 left breast, 39/5 right breast, and 30/5 prostate patients to auto-segment a total of 24 organs at risk (OARs). Subsequently, both vendor-pretrained and custom-trained models were tested on the remaining patients from each group. Auto-segmented contours were evaluated by comparing them with clinically approved contours via the Dice similarity coefficient (DSC) and mean surface distance (MSD). The performance of the left and right breast models was assessed jointly according to ipsilateral/contralateral locations. Results: The average DSCs for all structures in vendor-pretrained and custom-trained models were as follows: 0.81 ± 0.12 and 0.86 ± 0.11 in HN; 0.67 ± 0.16 and 0.80 ± 0.11 in the breast; and 0.87 ± 0.09 and 0.92 ± 0.06 in the prostate. The corresponding average MSDs were 0.81 ± 0.76 mm and 0.76 ± 0.56 mm (HN), 4.85 ± 2.44 mm and 2.42 ± 1.49 mm (breast), and 2.17 ± 1.39 mm and 1.21 ± 1.00 mm (prostate). Notably, custom-trained models showed significant improvements over vendor-pretrained models for 14 of 24 OARs, reflecting the influence of data/contouring variations in segmentation performance. Conclusions: These findings underscore the substantial impact of institutional preferences and clinical practices on the implementation of vendor-pretrained models. We also found that a relatively small amount of institutional data was sufficient to train customized segmentation models with sufficient accuracy.

CHUWD-H v1.0: a comprehensive historical hourly weather database for U.S. urban energy system modeling

Reliable and continuous meteorological data are crucial for modeling the responses of energy systems and their components to weather and climate conditions, particularly in densely populated urban areas. However, existing long-term datasets often suffer from spatial and temporal gaps and inconsistencies, posing great challenges for detailed urban energy system modeling and cross-city comparison under realistic weather conditions. Here we introduce the Historical Comprehensive Hourly Urban Weather Database (CHUWD-H) v1.0, a 23-year (1998–2020) gap-free and quality-controlled hourly weather dataset covering 550 weather station locations across all urban areas in the contiguous United States. CHUWD-H v1.0 synthesizes hourly weather observations from stations with outputs from a physics-based solar radiation model and a reanalysis dataset through a multi-step gap filling approach. A 10-fold Monte Carlo cross-validation suggests that the accuracy of this gap filling approach surpasses that of conventional gap filling methods. Designed primarily for urban energy system modeling, CHUWD-H v1.0 should also support historical urban meteorological and climate studies, including the validation and evaluation of urban climate modeling.

Spatial variation and associated factors of home delivery among reproductive age group women in Ethiopia, evidence from Performance Monitoring for Action Ethiopia Survey 2019, spatial and multilevel logistic regression analysis

IntroductionHome birth is described as a delivery that takes place at home without the presence of a skilled birth attendant. In 2017, nearly 295,000 mothers died from various pregnancy and childbirth-related problems, accounting for approximately 810 maternal deaths per day. Therefore, this study aims to investigate the spatial distributions of home birth and associated factors in Ethiopia using the Performance Monitoring for Action Survey (PMAS) 2019) to get information that helps to take geographic-based interventions and can assist health planners and policymakers in developing particular measures to reduce home deliveries.MethodIn PMA-ET 2019, a community-based cross-sectional study was conducted in collaboration with Addis Ababa University, Johns Hopkins University, and the Federal Ministry of Health from September 2019 to December 2019, in Ethiopia. A multi-stage cluster sampling procedure was employed to draw from the stratified 2019 PMAS sample. A weighted total of 5,796 women were included in this study. ArcGIS version 10.7 software was used to visualize the spatial analysis. In addition, STATA version 14 of the statistical software was used for multilevel analysis The Bernoulli model was applied using Kulldorff's SaTScan version 9.6 software to identify significant purely spatial clusters for home delivery in Ethiopia. Intra-class Correlation Coefficient (ICC), Likelihood Ratio (LR) test, Median Odds Ratio (MOR), and deviance (−2LLR) values were used for model comparison and fitness. Adjusted Odds Ratios (AOR) with a 95% Confidence Interval (CI) and p-value &amp;lt;0.05 in the multilevel logistic model were used to declare significant factors associated with home delivery.ResultThe spatial distribution of home delivery was non-random in Ethiopia. Statistically significant high hotspots of home delivery were found in Somali, Afar, Sidama, most of South Nation Nationality and People Region (SNNP), most parts of Amhara, south west Ethiopia, and Oromia region. In the multilevel logistic regression model; Women from the lowest wealth quintile were 1.68 times [AOR = 1.68; 95% CI: 1.31, 2.15] higher odds of giving birth at home as compared to their counterparts. Regarding maternal educational status, mothers who had no education, primary education, and secondary education had 9.91 times [AOR = 9.91, 95% CI: 5.44, 18.04], 6.62 times [AOR = 6.62, 95% CI: 3.65, 12.00] and 2.99 times [AOR = 2.99, 95% CI: 1.59, 5.63] higher odds of giving birth at home compared to mothers who attained higher education, respectively. In addition, community-level factors were significantly associated with home delivery, women who had high community-level poverty were 1.76 times [AOR = 1.76; 95% CI: 1.14, 2.72] higher odds of home delivery compared to women who had low community-level poverty.ConclusionHome delivery was statistically found to be a significantly high hot spot in Somalia, Afar, Sidama, most of the South Nation Nationality and People area (SNNP), most of Amhara, southwest Ethiopia, and the Oromia region of Ethiopia. Significant factors associated with home delivery in Ethiopia were women with lower levels of education, poor wealth, living in rural areas, high levels of community poverty, divorced or separated widowed marital status, and older maternal ages. Therefore, health institutions, health professionals, National and regional policymakers health planners community leaders and all concerned should give priority to the identified hot spot clusters to design an effective intervention program to reduce home delivery.

On the Benefit of Robust Bayesian Confirmatory Factor Analysis

Data normality is a typical assumption when performing confirmatory factor analysis. However, always using confirmatory factor analysis under the normally assumption will often lead to poor model fit in real-world applications. In this work, we construct an efficient sampler for a robust Bayesian confirmation factor analysis by utilizing the Student’s t distribution and provide a novel way to calculate the corresponding marginal likelihood. We demonstrate that the proposed strategy is both computationally efficient and performs favorably when compared to standard approaches that rely on the normality assumption. Finally, we provide a research toolkit for robust Bayesian confirmatory factor analysis and model comparison.

Using the exponential distribution as a synthetic curve volume of runoff to determine the volume of surface runoff

The paper analyzes the problem of determining the amount of surface wastewater that has formed on the territory of the drainage area as a result of precipitation and snowmelt, and enters the surface sewage disposal systems, for which the flow coefficient is used, which reflects the relationship between waterproofness and flow. Comparison of regression models used to estimate runoff coefficients showed a significant variation of its values ​​over the entire range of the total impervious area. For the samples of the generated values ​​of the total runoff coefficient with the corresponding proportion of waterproofness, statistical indicators were calculated and box diagrams were constructed. Regression equations of the total runoff coefficient from the share of the total impervious area with the corresponding soil filtration coefficient were obtained for the conditions of Kyiv, which more fully take into account the infiltration potential of the soil and the moisture level of the catchment before the rain starts.

SELECTING WELDING SAFETY GOGGLES BY MULTI-CRITERIA DECISION-MAKING APPROACHES— A COMPREHENSIVE EVALUATION

This paper presents a comprehensive application of the Multi-Criteria Decision-Making (MCDM) approach to evaluate and rank different models of welding goggles. The study uses both the Mean Weight and Entropy Weight methods in conjunction with five different MCDM techniques—TOPSIS, CODAS, COPRAS, SAW, and MOORA. These methods provide an insightful comparison and robust ranking of various goggle models. The analysis consistently highlighted 'ArcOne The Fly Safety Goggles' and '3M Speedglas 100 Series' as superior performers across multiple criteria.

How penetration of e-commerce in rural areas drives consumption in China

PurposeThis paper explores how rural e-commerce drives consumption in both urban and rural residents; further, it presents evidence from the perspective of e-commerce penetration in rural areas and among farmers to assess the contribution of e-commerce to China’s economic strategy goal of fostering an economy driven by domestic demand.Design/methodology/approachData of Taobao villages from 2014 to 2021 are matched with data on prefecture-level cities. A two-way fixed effect model, instrumental variable model and group comparison method are employed to explore the role and mechanism of e-commerce penetration in rural areas in stimulating residents’ consumption through consumption propensity, purchasing power and consumer population.FindingsThe findings confirm that e-commerce penetration of rural areas can stimulate residential consumption, and residents generally adhere to the consumption concept of “living within their means.” The driving effect on residents’ consumption is achieved by expanding the urban consumer population and improving the purchasing powers of both urban and rural residents. The spatial dispersion of e-commerce villages strengthens the positive impact of e-commerce penetration in rural areas on consumption; this strengthening is caused by the increasing consumer population of urban residents and the purchasing power of both urban and rural residents. The agriculture-oriented e-commerce penetration of rural areas negatively impacts residents’ consumption, consumption propensity and consumer population; however, industrial-oriented penetration positively affects residents’ consumption, consumption propensity, purchasing power and consumer population.Originality/valueTo the best of our knowledge, this paper is the first to analyze the role and mechanism of e-commerce penetration in rural areas in stimulating residents’ consumption through consumption propensity, purchasing power and consumer population. This study contributes to advancing e-commerce and consumer theory by exploring the realm of rural e-commerce while also presenting empirical evidence from China.

Empirical Comparison of Facebook Prophet and Traditional Models for Tomato Price Forecasting in Greece

Agricultural product prices are crucial to the income and livelihoods of millions worldwide, making accurate forecasting essential for market participants. The European Union regularly reports agricultural product prices through an online databank, ensuring up-to-date price information is widely available. This study evaluates the effectiveness of Facebook Prophet, a modern forecasting tool, in predicting tomato prices in Greece from 2013 to 2024.The results show that Facebook Prophet outperforms traditional forecasting models, providing more accurate predictions with lower Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE) values. Notably, the model exhibited superior performance in forecasting tomato prices over a six-month horizon compared to conventional seasonally adjusted models. This demonstrates Facebook Prophet’s potential to significantly improve decision-making in agricultural markets, offering reliable price forecasts to stakeholders such as farmers, traders, and policymakers. Furthermore, the study incorporated feedback from market participants, which provided valuable insights into market practices and conditions. This integration of practical knowledge with advanced forecasting techniques enhanced the interpretation of the results, making them more applicable to real-world scenarios. Overall, the findings suggest that Facebook Prophet holds considerable promise for future agricultural price forecasting, with potential applications across various commodities. These insights pave the way for more precise agricultural forecasts, benefiting all market participants by supporting more informed and timely decision-making.

Short-Term Energy Generation Forecasts at a Wind Farm—A Multi-Variant Comparison of the Effectiveness and Performance of Various Gradient-Boosted Decision Tree Models

High-quality short-term forecasts of wind farm generation are crucial for the dynamically developing renewable energy generation sector. This article addresses the selection of appropriate gradient-boosted decision tree models (GBDT) for forecasting wind farm energy generation with a 10-min time horizon. In most forecasting studies, authors utilize a single gradient-boosted decision tree model and compare its performance with other machine learning (ML) techniques and sometimes with a naive baseline model. This paper proposes a comprehensive comparison of all gradient-boosted decision tree models (GBDTs, eXtreme Gradient Boosting (XGBoost), Light Gradient-Boosting Machine (LightGBM), and Categorical Boosting (CatBoost)) used for forecasting. The objective is to evaluate each model in terms of forecasting accuracy for wind farm energy generation (forecasting error) and computational time during model training. Computational time is a critical factor due to the necessity of testing numerous models with varying hyperparameters to identify the optimal settings that minimize forecasting error. Forecast quality using default hyperparameters is used here as a reference. The research also seeks to determine the most effective sets of input variables for the predictive models. The article concludes with findings and recommendations regarding the preferred GBDT models. Among the four tested models, the oldest GBDT model demonstrated a significantly longer training time, which should be considered a major drawback of this implementation of gradient-boosted decision trees. In terms of model quality testing, the lowest nRMSE error was achieved by the oldest model—GBDT in its tuned version (with the best hyperparameter values obtained from exploring 40,000 combinations).

Relating visual and pictorial space: Integration of binocular disparity and motion parallax.

Traditionally, perceptual spaces are defined by the medium through which the visual environment is conveyed (e.g., in a physical environment, through a picture, or on a screen). This approach overlooks the distinct contributions of different types of visual information, such as binocular disparity and motion parallax, that transform different visual environments to yield different perceptual spaces. The current study proposes a new approach to describe different perceptual spaces based on different visual information. A geometrical model was developed to delineate the transformations imposed by binocular disparity and motion parallax, including (a) a relief depth scaling along the observer's line of sight and (b) pictorial distortions that rotate the entire perceptual space, as well as the invariant properties after these transformations, including distance, three-dimensional shape, and allocentric direction. The model was fitted to the behavioral results from two experiments, wherein the participants rotated a human figure to point at different targets in virtual reality. The pointer was displayed on a virtual frame that could differentially manipulate the availability of binocular disparity and motion parallax. The model fitted the behavioral results well, and model comparisons validated the relief scaling in the form of depth expansion and the pictorial distortions in the form of an isotropic rotation. Fitted parameters showed that binocular disparity renders distance invariant but also introduces relief depth expansion to three-dimensional objects, whereas motion parallax keeps allocentric direction invariant. We discuss the implications of the mediating effects of binocular disparity and motion parallax when connecting different perceptual spaces.

Active and Passive Data Combined Depth Inversion Based on Multi-Temporal Observation: Comparison of Model and Strategy

Active and Passive Data Combined Depth Inversion Based on Multi-Temporal Observation: Comparison of Model and Strategy

Prediction Method for Pomegranate Chlorophyll Content Based on Multi-feature Fusion of Unmanned Aerial Vehicle

The aim of this study was to obtain RGB and multispectral images of fruit tree canopy during the flowering period of pomegranate by multispectral unmanned aerial vehicle (UAV) to quickly and accurately predict the chlorophyll content in order to improve the monitoring efficiency of the orchard. A handheld chlorophyll meter was used to obtain the actual chlorophyll values, and image processing techniques were combined to extract parameters such as color features and texture features of the RGB images as well as vegetation index of the multispectral images. A chlorophyll content prediction method based on the support vector regression model and the convolutional neural network model (CNN-Attention) combined with the attention mechanism was established. The results showed that (1) the accuracy of the prediction model was improved by the fusion of RGB image features and multispectral image vegetation index. (2) After model comparison, the CNN-Attention model built under the fused features was the best in chlorophyll content prediction with R2, RE, and RMSE of 0.9699, 0.0052, and 0.6013, respectively.This study provides a more accurate method for fruit tree chlorophyll content prediction using UAVs, which provides a practical reference for orchard management.

A benchmark comparison of AI-based modeling of soil infiltration rates

ABSTRACT Infiltration is crucial in the hydrological cycle, serving as the primary process that increases soil moisture. This study investigates the estimation of soil infiltration rate (IR) using various techniques, including the group method of data handling, Gaussian process, support vector machine, artificial neural network (ANN), and multivariate adaptive regression splines. A total of 190 field observations were collected from Alashtar sub-watersheds in Lorestan, Iran. About 70% of the observations were used for model preparation, while 30% were used for validation. The input variables for the study are time, sand, clay, silt, pH, electrical conductivity, moisture content, soil bulk density, porosity, calcium carbonate, phosphorus, organic carbon, organic matter, nitrogen, and temperature, while IR is the output variable. Obtained results indicate that the ANN has a higher accuracy with coefficient of correlation values of 0.9366, 0.8624, mean absolute error values as 0.0607, 0.1000, Nash–Sutcliffe model efficiency values of 0.8732, 0.7350, the scattering index values of 0.3108, 0.5003, and the Legates and McCabe's index values of 0.6585, 0.5654 by using training and testing datasets, respectively. A sensitivity analysis highlighted that time is the parameter that most influences estimating the IR. The study underscores the precision of ANN in predicting soil IRs and the need for soft computing models in hydrological models to improve accuracy and reliability in IR prediction.

Comparison of tracer kinetic models in differentiating malignant from normal prostate tissue using dynamic contrast-enhanced MRI

PurposeThe aim of this study was to evaluate the diagnostic value of dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) derived kinetic parameters with high spatiotemporal resolution in discriminating malignant from normal prostate tissues.MethodsFifty patients with suspicious of malignant diseases in prostate were included in this study. Regions of interest (ROI) were manually delineated by experienced radiologists. Voxel-wise kinetic parameters were produced with the following tracer kinetic models (TKMs): Tofts model, extended Tofts model (ETM), Brix’s conventional two-compartment model (Brix), adiabatic tissue homogeneity model (ATH), and distributed parameter model (DP). The initial area under the signal-time curve (IAUC) with an uptake integral approach was also included. Mann–Whitney U test and receiver operating characteristic (ROC) curves were used to evaluate the capability of distinguishing tumor lesions from normal tissues. A p-value of 0.05 or less is considered statistically significant. ROI based parameters correlation analysis between DP and ETM were performed.Results624 lesions and 269 normal tissue ROIs were obtained. Thirty parameters were derived from the six kinetic models. Except for PS from Brix, statistically significant differences between lesions and normal tissues (P&amp;lt;0.05) were observed in other parameters.Ve from DP, ATH and Brix and PS from ATH have AUC values less than 0.6 in the ROC analysis. MTT, Vp and PS from DP, Ktrans from ETM and Tofts, E and PS from ATH, IAUC parameters and F from Brix have AUC values larger than 0.8. Ve and Vp from DP and ETM are correlated (r&amp;gt; 0.65). The correlation coefficient between Ktrans from ETM and PS from DP is 0.751.ConclusionMTT, Vp and PS from DP, Ktrans from ETM and Tofts, E and PS from ATH, F from Brix and IAUC parameters can be used to differentiate malignant lesions from normal tissues in the prostate.

Simplified Scan-vs-BIM Frameworks for Automated Structural Inspection of Steel Structures

This paper presents a deep learning-based Scan-vs-BIM methodology for evaluating structural integrity through the extraction of features from As-Built scan and As-Planned Building Information Modeling (BIM) comparison data. Traditional Scan-vs-BIM frameworks often rely on Scan-to-BIM processes to generate point cloud-based mesh models for comparison, which significantly impairs computational efficiency. In contrast, the proposed streamlined Scan-vs-BIM framework incorporates a deep neural network (DNN) model consisting of two neural networks: one for structural integrity assessment and another for error type analysis. The model evaluates the structural integrity of individual components in a sequential manner, repeating the process across all elements to comprehensively assess the entire structure. Rather than converting point cloud data into mesh models for comparison, this approach directly measures the spatial discrepancies between the As-Built point cloud and As-Planned BIM, analyzing the distribution tendencies of these distance values. Experimental validation on actual steel structures demonstrated that the proposed method effectively predicts structural integrity, providing significant improvements in both accuracy and computational performance.

Effect of radiobiological parameters on the TCP for breast cancer radiotherapy

Breast cancer remains the most prevalent malignancy affecting women globally. Among the various treatment modalities, radiotherapy stands out as a cornerstone for tumor eradication. This research explores the impact of radiosensitivity parameters on Tumor Control Probability (TCP) in breast cancer, with an emphasis on distinct radiotherapeutic techniques such as conventional, hypofractionated, and FAST, as well as the role of tumor repopulation. Based on the literature review, we obtained data on α and β radiosensitivity parameters, cell repopulation rates and standard breast cancer treatment protocols. These parameters informed the calculation of the fraction of cells surviving irradiation via the linear-quadratic model, facilitating an assessment of treatment efficacy through the Poissonian TCP model. Our findings underscore the critical influence of radiosensitivity parameters α and β on treatment outcomes, with β emerging as the predominant factor due to its quadratic contribution to the survival fraction. Moreover, our analysis indicates that tumor growth is negligible relative to the substantial cell mortality induced by radiation in the case of breast cancer. Techniques such as FAST and hypofractionated radiotherapy were identified as particularly effective, offering expedited tumor control, especially with elevated α and β values. The quadratic term β significantly enhances treatment success, while tumor repopulation exerts minimal influence on TCP, corroborating previous model comparisons. Notably, higher doses per fraction, rather than increased cumulative doses, were associated with improved TCP, providing a critical insight for optimizing radiotherapy protocols. Currently, radiobiology is not systematically integrated into clinical practice, and its analysis through PCT optimizes radiotherapy treatments, improving patient quality of life and healthcare delivery.

A multi-view prognostic model for diffuse large B-cell lymphoma based on kernel canonical correlation analysis and support vector machine

Background and objectivePositron emission tomography/computed tomography (PET/CT) is recommended as the standard imaging modality for diffuse large B-cell lymphoma (DLBCL) staging. However, many studies have neglected the role of patients' prognostic factors with respect to imaging PET/CT of quantitative features. In this paper, a multi-view learning (MVL) model is established to make full use of both clinical and imaging data to predict the prognosis of DLBCL patients and thereby assist doctors in decision-making.MethodsFeature engineering, including feature extraction, feature screening by recursive feature elimination, and dimensionality reduction by principal component analysis, are successively performed on the clinical data and imaging data of the research subjects to obtain the study data. After dividing the data into training and test sets, an instance weighting method is applied to the training data. Subsequently, kernel mapping is performed on the imaging features and clinical features separately, and this kernel mapping is processed in the new kernel feature space using kernel canonical correlation analysis (KCCA). Lastly, model training is performed on the obtained common kernel subspace using a support vector machine (SVM). The final overall model, named SVM-2view-KCCA (SVM-2 K), was compared with three other multi-view models (Ensemble-SVM, Multi-view maximum entropy discrimination, and canonical correlation analysis). The performance of the model was evaluated on the test data with respect to several dichotomous metrics: accuracy, sensitivity, F1 score, the area under the curve (AUC), and G-mean.ResultsThe SVM model improved AUC by 10.5%, sensitivity by 11.9%, accuracy by 9.8%, F1 score by 9.2%, and G-mean by 7.8% for the DLBCL test data after feature engineering based on dimensionality reduction and instance weighting. In the performance comparison of single-view learning models, the SVM-based integration of clinical and imaging features achieved the best overall performance (AUC = 86.3%, accuracy = 91.6%, sensitivity = 83.2%, F1 = 85.7%, and G-mean = 86.1%). In the comparison of MVL models, SVM-2 K achieved the best overall performance (AUC = 92.1%, accuracy = 96.9%, sensitivity = 90.9%, F1 = 92.8%, and G-mean = 91.4%), and the performance of each MVL model was better than that of the best single-view learning model.ConclusionsMVL models outperformed single-view learning models. Of the MVL models, the proposed SVM-2 K achieved the best overall performance and could accurately predict patient prognosis.

The size of myocardial infarction and peri-infarction edema are not major determinants of diastolic impairment after acute myocardial infarction.

To study the relationship between myocardial infarction size (IS), myocardial edema, and diastolic dysfunction after acute myocardial infarction (MI) both in the acute phase, and in the development of diastolic dysfunction in the follow-up setting. A further purpose is to study diastolic function using a mechanistic model as well as conventional parameters. Patients underwent cardiovascular magnetic resonance (CMR) imaging and echocardiography including mechanistic analysis using the parameterized diastolic filling method within 4-7days (acute) and 6months after a first acute anterior MI (n = 74). Linear regression modeling of echocardiographic diastolic parameters using CMR IS with and without inclusion of the myocardium at risk (MAR) and model comparisons with likelihood ratio tests were performed. Diastolic parameters at 6months follow-up were modelled using final IS. For most parameters there was no association with acute IS, except for deceleration time (R2 = 0.24, p < 0.001), left atrial volume index (R2 = 0.13, p = 0.01) and the mechanistic stiffness parameter (R2 = 0.21, p < 0.001). Adding MAR improved only the e' model (adjusted R2 increase: 0.08, p = 0.02). At 6months follow-up, final IS was only associated with viscoelastic energy loss (R2 = 0.22, p = 0.001). In acute MI, both IS and MAR are related to diastolic function but only to a limited extent. At 6months after infarction, increasing IS is related to less viscoelastic energy loss, albeit also to a limited extent. The relationship between IS and diastolic dysfunction seems to be mediated by mechanisms beyond simply the spatial extent of ischemia or infarction.