Entire Dataset Research Articles

Unsupervised Feature Selection to Identify Important ICD-10 and ATC Codes for Machine Learning on a Cohort of Patients With Coronary Heart Disease: Retrospective Study.

The application of machine learning in health care often necessitates the use of hierarchical codes such as the International Classification of Diseases (ICD) and Anatomical Therapeutic Chemical (ATC) systems. These codes classify diseases and medications, respectively, thereby forming extensive data dimensions. Unsupervised feature selection tackles the "curse of dimensionality" and helps to improve the accuracy and performance of supervised learning models by reducing the number of irrelevant or redundant features and avoiding overfitting. Techniques for unsupervised feature selection, such as filter, wrapper, and embedded methods, are implemented to select the most important features with the most intrinsic information. However, they face challenges due to the sheer volume of ICD and ATC codes and the hierarchical structures of these systems. The objective of this study was to compare several unsupervised feature selection methods for ICD and ATC code databases of patients with coronary artery disease in different aspects of performance and complexity and select the best set of features representing these patients. We compared several unsupervised feature selection methods for 2 ICD and 1 ATC code databases of 51,506 patients with coronary artery disease in Alberta, Canada. Specifically, we used the Laplacian score, unsupervised feature selection for multicluster data, autoencoder-inspired unsupervised feature selection, principal feature analysis, and concrete autoencoders with and without ICD or ATC tree weight adjustment to select the 100 best features from over 9000 ICD and 2000 ATC codes. We assessed the selected features based on their ability to reconstruct the initial feature space and predict 90-day mortality following discharge. We also compared the complexity of the selected features by mean code level in the ICD or ATC tree and the interpretability of the features in the mortality prediction task using Shapley analysis. In feature space reconstruction and mortality prediction, the concrete autoencoder-based methods outperformed other techniques. Particularly, a weight-adjusted concrete autoencoder variant demonstrated improved reconstruction accuracy and significant predictive performance enhancement, confirmed by DeLong and McNemar tests (P<.05). Concrete autoencoders preferred more general codes, and they consistently reconstructed all features accurately. Additionally, features selected by weight-adjusted concrete autoencoders yielded higher Shapley values in mortality prediction than most alternatives. This study scrutinized 5 feature selection methods in ICD and ATC code data sets in an unsupervised context. Our findings underscore the superiority of the concrete autoencoder method in selecting salient features that represent the entire data set, offering a potential asset for subsequent machine learning research. We also present a novel weight adjustment approach for the concrete autoencoders specifically tailored for ICD and ATC code data sets to enhance the generalizability and interpretability of the selected features.

Construction of a nomogram with IrAE and clinic character to predict the survival of advanced G/GEJ adenocarcinoma patients undergoing anti-PD-1 treatment.

This study aimed to develop and validate a survival prediction model and nomogram to predict survival in patients with advanced gastric or gastroesophageal junction (G/GEJ) adenocarcinoma undergoing treatment with anti-programmed cell death 1 receptor (PD-1). This model incorporates immune-related adverse events (irAEs) alongside common clinical characteristics as predictive factors. A dataset comprising 255 adult patients diagnosed with advanced G/GEJ adenocarcinoma was assembled. The irAEs affecting overall survival (OS) to a significant degree were identified and integrated as a candidate variable, together with 12 other candidate variables. These included gender, age, Eastern cooperative oncology group performance status (ECOG PS) score, tumor stage, human epidermal growth factor receptor 2 (HER2) expression status, presence of peritoneal and liver metastases, year and line of anti-PD-1 treatment, neutrophil-to-lymphocyte ratio (NLR), controlling nutritional status (CONUT) score, and Charlson comorbidity index (CCI). To mitigate timing bias related to irAEs, landmark analysis was employed. Variable selection was performed using the least absolute shrinkage and selection operator (LASSO) regression to pinpoint significant predictors, and the variance inflation factor was applied to address multicollinearity. Subsequently, a Cox regression analysis utilizing the forward likelihood ratio method was conducted to develop a survival prediction model, excluding variables that failed to satisfy the proportional hazards (PH) assumption. The model was developed using the entire dataset, then internally validated through bootstrap resampling and externally validated with a cohort from another Hospital. Furthermore, a nomogram was created to delineate the predictive model. After consolidating irAEs from the skin and endocrine systems into a single protective irAE category and applying landmark analysis, variable selection was conducted for the prognostic prediction model along with other candidate variables. The finalized model comprised seven variables: ECOG PS score, tumor stage, HER2 expression status in tumor tissue, first-line anti-PD-1 treatment, peritoneal metastasis, CONUT score, and protective irAE. The overall concordance index for the model was 0.66. Calibration analysis verified the model's accuracy in aligning predicted outcomes with actual results. Clinical decision curve analysis indicated that utilizing this model for treatment decisions could enhance the net benefit regarding 1- and 2-year survival rates for patients. This study developed a prognostic prediction model by integrating common clinical characteristics of irAEs and G/GEJ adenocarcinoma. This model exhibits good clinical practicality and possesses accurate predictive ability for overall survival OS in patients with advanced G/GEJ adenocarcinoma.

Impacts of future climate change on rice yield based on crop model simulation—A meta-analysis

Rice is one of the world's major food crops. Changes in major climatic factors such as temperature, rainfall, solar radiation and carbon dioxide (CO2) concentration have an important impact on rice growth and yield. However, many of the current studies that predict the impact of future climate change on rice yield are affected by uncertainties such as climate models, climate scenarios, model parameters and structure, and showing great differences. This study was based on the assessment results of the impact of climate change on rice in the future of 111 published literature, and comprehensively analyzed the impact and uncertainty of climate change on rice yield. This study utilized local polynomial (Loess) regression analysis to investigate the impact of changes in mean temperature, minimum temperature, maximum temperature, solar radiation, and precipitation on relative rice yield variations within a complete dataset. A linear mixed-effects model was used to quantitatively analyze the relationships between the restricted datasets. The qualitative analysis based on the entire dataset revealed that rice yields decreased with increasing average temperature. The precipitation changed between 0 and 25 %, it was conducive to the stable production of rice, and when the precipitation changed >25 %, it would cause rice yield reduction. The change of solar radiation was less than −1.15 %, the rice yield increases with the increase of solar radiation, and when the change of solar radiation exceeds −1.15 %, the rice yield decreases. Elevated CO2 concentrations and management practices could mitigate the negative effects of climate change. The results of a quantitative analysis utilizing the mixed effects model revealed that average temperature, precipitation, CO2 concentration, and adaptation methods all had a substantial impact on rice production, and elevated CO2 concentrations and management practices could exert positive influences on rice production. For every 1 °C and 1 % increase in average temperature and precipitation, rice yield decreased by 3.85 % and 0.56 %, respectively. For every 100 ppm increase in CO2 concentration, rice yield increased by 7.1 %. The variation of rice yield under different climate models, study sites and climate scenarios had significant variability. Elevated CO2 concentrations and management practices could compensate for the negative effects of climate change, benefiting rice production. This study comprehensively collected and analyzed a wide range of literature and research, which provides an in-depth understanding of the impacts of climate change on rice production and informs future research and policy development.

Construction of a Multi-Label Classifier for Extracting Multiple Incident Factors From Medication Incident Reports in Residential Care Facilities: Natural Language Processing Approach.

Medication safety in residential care facilities is a critical concern, particularly when nonmedical staff provide medication assistance. The complex nature of medication-related incidents in these settings, coupled with the psychological impact on health care providers, underscores the need for effective incident analysis and preventive strategies. A thorough understanding of the root causes, typically through incident-report analysis, is essential for mitigating medication-related incidents. We aimed to develop and evaluate a multilabel classifier using natural language processing to identify factors contributing to medication-related incidents using incident report descriptions from residential care facilities, with a focus on incidents involving nonmedical staff. We analyzed 2143 incident reports, comprising 7121 sentences, from residential care facilities in Japan between April 1, 2015, and March 31, 2016. The incident factors were annotated using sentences based on an established organizational factor model and previous research findings. The following 9 factors were defined: procedure adherence, medicine, resident, resident family, nonmedical staff, medical staff, team, environment, and organizational management. To assess the label criteria, 2 researchers with relevant medical knowledge annotated a subset of 50 reports; the interannotator agreement was measured using Cohen κ. The entire data set was subsequently annotated by 1 researcher. Multiple labels were assigned to each sentence. A multilabel classifier was developed using deep learning models, including 2 Bidirectional Encoder Representations From Transformers (BERT)-type models (Tohoku-BERT and a University of Tokyo Hospital BERT pretrained with Japanese clinical text: UTH-BERT) and an Efficiently Learning Encoder That Classifies Token Replacements Accurately (ELECTRA), pretrained on Japanese text. Both sentence- and report-level training were performed; the performance was evaluated by the F1-score and exact match accuracy through 5-fold cross-validation. Among all 7121 sentences, 1167, 694, 2455, 23, 1905, 46, 195, 1104, and 195 included "procedure adherence," "medicine," "resident," "resident family," "nonmedical staff," "medical staff," "team," "environment," and "organizational management," respectively. Owing to limited labels, "resident family" and "medical staff" were omitted from the model development process. The interannotator agreement values were higher than 0.6 for each label. A total of 10, 278, and 1855 reports contained no, 1, and multiple labels, respectively. The models trained using the report data outperformed those trained using sentences, with macro F1-scores of 0.744, 0.675, and 0.735 for Tohoku-BERT, UTH-BERT, and ELECTRA, respectively. The report-trained models also demonstrated better exact match accuracy, with 0.411, 0.389, and 0.399 for Tohoku-BERT, UTH-BERT, and ELECTRA, respectively. Notably, the accuracy was consistent even when the analysis was confined to reports containing multiple labels. The multilabel classifier developed in our study demonstrated potential for identifying various factors associated with medication-related incidents using incident reports from residential care facilities. Thus, this classifier can facilitate prompt analysis of incident factors, thereby contributing to risk management and the development of preventive strategies.

Firearm ownership and suicide: Has the time come to discuss uniformity of health and social assessments in aid of regulation? Reflection from a retrospective study on a forensic case series.

Firearm-related deaths are an issue of ongoing public interest, from a health and economic perspective. Worldwide, firearm prevalence in suicides varies widely, depending on multiple factors including the availability of weapons in each country. Although several studies have shown that national laws about gun ownership, reducing legal access to guns, decrease the risk of suicide by firearm, the current situation clashes with widely differing legislations. Retrospective study on a forensic case series. Autopsy reports assessed at the Section of Legal Medicine of Milan (Italy) from January 2014 to December 2023 were retrospectively documented. Only firearm suicides were considered. For each case, a close analysis of the criminal offence reports has been performed to obtain information about the gender and age of the victim, as well as the legal possession of firearms, psychiatric disorders, alcohol, and illicit drug abuse. Among all the 1,164 suicides assessed at the Section of Legal Medicine of Milan over a 10-year-period, 101 cases (8.7%) were firearm-related. The male to female ratio was therefore about 13:1. No seasonal trends were observed. Most of the individuals owned the firearm license. Of the entire dataset, 35.6% suffered from psychiatric disorders, 4% of alcohol abuse and 2% of drug addiction. Among the owners of a firearm license, 42.3% had psychiatric disorders. Knowledge about firearm suicide and its relation to firearm legal possession is limited and current preventive laws should be reconsidered. Present administrative and clinical examinations required to obtain a firearm license in Italy and in some European territories are dealt with. The evaluation of the firearm-related risk of abuse is an essential but complex procedure, which requires not only clinic-anamnestic data but also in-deep psychiatric information. There is a need to develop and reinvigorate a debate that currently presents very heterogeneous solutions, but which would probably benefit from a common vision of the prevention strategies that can be implemented and enacted for the benefit of the entire community.

Advancing bladder cancer management: development of a prognostic model and personalized therapy.

Bladder cancer (BLCA) was recognized as a significant public health challenge due to its high incidence and mortality rates. The influence of molecular subtypes on treatment outcomes was well-acknowledged, necessitating further exploration of their characterization and application. This study was aimed at enhancing the understanding of BLCA by mapping its molecular heterogeneity and developing a robust prognostic model using single-cell and bulk RNA sequencing data. Additionally, immunological characteristics and personalized treatment strategies were investigated through the risk score. Single-cell RNA sequencing (scRNA-seq) data from GSE135337 and bulk RNA-seq data from several sources, including GSE13507, GSE31684, GSE32894, GSE69795, and TCGA-BLCA, were utilized. Molecular subtypes, particularly the basal-squamous (Ba/Sq) subtype associated with poor prognosis, were identified. A prognostic model was constructed using LASSO and Cox regression analyses focused on genes linked with the Ba/Sq subtype. this model was validated across internal and external datasets to ensure predictive accuracy. High- and low-risk groups based on the risk score derived from TCGA-BLCA data were analyzed to examine their immune-related molecular profiles and treatment responses. Six molecular subtypes were identified, with the Ba/Sq subtype being consistently associated with poor prognosis. The prognostic model, based on basal-squamous subtype-related genes (BSSRGs), was shown to have strong predictive performance across diverse clinical settings with AUC values at 1, 3, and 5 years indicating robust predictability in training, testing, and entire datasets. Analysis of the different risk groups revealed distinct immune infiltration and microenvironments. Generally higher tumor mutation burden (TMB) scores and lower tumor immune dysfunction and exclusion (TIDE) scores were exhibited by the low-risk group, suggesting varied potentials for systemic drug response between the groups. Finally, significant differences in potential systemic drug response rates were also observed between risk groups. The study introduced and validated a new prognostic model for BLCA based on BSSRGs, which was proven effective in prognosis prediction. The potential for personalized therapy, optimized by patient stratification and immune profiling, was highlighted by our risk score, aiming to improve treatment efficacy. This approach was promised to offer significant advancements in managing BLCA, tailoring treatments based on detailed molecular and immunological insights.

Diagnostic criteria for eosinophilic chronic rhinosinusitis: Comparative analysis and novel scoring system.

Accurate identification of eosinophilic chronic rhinosinusitis is essentialg because its treatment and prognosis substantially differ from other subtypes. This retrospective observational study included 640 patients who underwent endoscopic sinus surgery for chronic rhinosinusitis in a single tertiary center from January 2021 to December 2022. Receiver operating characteristic curves were generated to compare accuracy, sensitivity, specificity of the novel scoring system, and previous diagnostic criteria (Japanese Epidemiological Survey of Refractory Eosinophilic Chronic Rhinosinusitis, European Forum for Research and Education in Allergy and Airway Diseases, European Position Paper on Rhinosinusitis and Nasal Polyps, and Sakuma etal.) for predicting eosinophilic chronic rhinosinusitis (ECRS) by tissue eosinophil count ≥70 per high power field. Patients were randomly divided into estimation (n=430) and validation (n=210) groups. The area under the receiver operating characteristic curve for the novel score was 0.753 (95% confidence interval [CI], 0.670-0.835) in the estimation group, 0.729 (0.629-0.830) in the validation group, and 0.661 (0.584-0.738) in the 20-fold cross-validation with the entire dataset. We propose a novel scoring system that incorporates three key parameters: "novel score=blood eosinophil (%) + total Lund-Mackay score of anterior ethmoid sinuses + 2 if nasal polyp present" greater than 7 can be reliably used for diagnosing ECRS. This system can facilitate decision-making processes regarding the administration of oral steroids and biologics targeting type 2 inflammation prior to surgical intervention.

The missing economies of density and south Australian local government efficiency

ABSTRACT Sound theoretical grounds exist to assume that economies of density will have pervasive effects on municipal service provision. However, to date the available empirical literature offers inconsistent and confusing findings on the potential impact of density. Moreover, the extant empirical evidence of substantial density economies is unconvincing. This paper investigates this apparent contradiction by addressing (a) significant non-monotonicity in density economies, (b) confusion over density economies through its disparate impact on various municipal activities and (c) underlying differences in how density affects different types of local council. These questions are investigated employing multiple regression analysis using a panel data set covering all 68 South Australian local authorities over the seven-year period from 2014 to 2020. We find evidence of economies of density across the entire data set. We infer that the measurement of population density imposes constraints on identifying density economies and the empirical methodology employed.

Geochemical indices and sedimentation rates in the Pardo River basin, São Paulo state, Brazil

Sediment geochemical indices and sedimentation rates data from the Pardo River watershed, São Paulo State, Brazil, provide relevant reference guidelines for stakeholders to plan future actions aiming to achieve the appropriate management of that hydrographic basin. The watershed includes 20 municipalities where about 650,000 inhabitants are living and whose waters after withdrawal and treatment supply ∼300,000 people in that region. Six sediment cores provided the geochemical and radiometric dataset, which was obtained from sampling sites located in the following cities: Ourinhos, Santa Cruz do Rio Pardo, Águas de Santa Bárbara, Avaré, Botucatu, and Pardinho. The concentration data for some metals were compared with the reference values established by the Brazilian environmental legislation for sediment dredging because currently there is a lack of specific quality guidelines for marine or fluvial sediment in Brazil. Also, the entire dataset was used in the calculation of two traditional geochemical indices, i.e., the enrichment factor (EF) and contamination factor (CF). In some cores, both indices provided insight revealing possible anthropogenic inputs of magnesium oxide (MgO), potassium oxide (K2O), sulfur trioxide (SO3), vanadium oxide (V2O5), copper (II) oxide (CuO), and other constituents due to development of activities related to the crop production. The constant rate of supply (CRS) of the unsupported/excess lead 210 (210Pb) model has been successfully applied to the obtained a radiometric 210Pb dataset, yielding mean values of the mass accumulation rate (MAR) between 0.86 and 7.23 g/(cm2·y) and linear sedimentation rates (LSR) ranging from 2.9 to 7.1 cm/y. Both of these rates exhibited a significant correlation, representing physical weathering processes occurring in the basin. Chemical weathering rates reported in the literature indicate values much lower than the physical weathering rates and such findings have also been confirmed in the current study when the obtained data were compared with results from previous investigations in the Pardo River watershed. The CRS 210Pb chronological model was useful to track historical changes in the MgO and CuO concentration in the sediment cores, allowing curves to be plotted exhibiting major peaks in core 1 (1980–2010) and core 3 (2000–2010), a continuous increase from 1970 onwards until 2010 (core 4), and a sharp increase at the end of the monitoring period, 2017 (cores 5 and 6). Data available for the harvested area at a municipality located within the basin for the period 1990–2022 have indicated a more pronounced increase over the last years in the production of sugar cane, soybeans, and oranges, implying an increasing use of fertilizers and agricultural correctives for crop production, whose chemical composition would justify the trends observed for MgO and CuO concentrations in the sediment cores. The current study also revealed a wide range of significant relations involving the concentration of ferric oxide (Fe2O3) and other analyzed constituents, with iron tending to accumulate in the finer grain sizes (from 0.125 to <0.037 mm) of the sediment. The CRS 210Pb chronological model permitted evaluation of the history of the iron flux up to the present, which could represent baseline guidance levels for indicating future anthropic inputs in the basin.

Repurposing of FDA-approved drugs against oligomerization domain of dengue virus NS1 protein: a computational approach.

Dengue fever is a serious health hazard on a global scale and its primary causative agent is the dengue virus (DENV). The non-structural protein 1 (NS1) of DENV plays a pivotal role in pathogenesis. It is associated with several autoimmune events, endothelial cell apoptosis, and vascular leakage, which increase mainly during the critical phase of infection. In this study, important residues of the oligomerization domain of NS1 protein were identified by literature searches. Virtual screening has been conducted using the entire dataset of the DrugBank database and the potential small-molecule inhibitors against the NS1 protein have been chosen on the basis of binding energy values. This is succeeded by molecular dynamics (MD) simulations of the shortlisted compounds, ultimately giving rise to five compounds. These five compounds were further subjected to RAMD simulations by applying a random direction force of specific magnitude on the ligand center of mass in order to push the ligand out of the protein-binding pocket, for the quantitative estimation of their binding energy values to determine the interaction strength between protein and ligand which prevents ligand unbinding from its binding site, ultimately leading to the selection of three major compounds, DB00826 (Natamycin), DB11274 (Dihydro-alphaergocryptine), and DB11275 (Epicriptine), with the DB11274 having a role against idiopathic Parkinson's disease, and thus may have possible important roles in the prevention of dengue-associated Parkinsonism. These compounds may act as prospective drugs against dengue, by preventing the oligomerization of the NS1 protein, thereby preventing disease progression and pathogenesis.

A Time-Dependent Damage Constitutive Model for Brittle Rock Materials Based on Subcritical Damage Theory.

In the domain of geotechnical engineering, a profound understanding of the long-term mechanical deformation characteristics of rocks is indispensable for the design and construction of structures, dams, tunnels, and various engineering projects. The deformation behavior of rocks under long-term loads directly impacts the stability and safety of engineering structures. This study employs micromechanical methods to investigate the subcritical extension of microcracks under stress corrosion. By examining the accumulated damage resulting from this phenomenon, the research explores the patterns of aging damage development and establishes a constitutive model for aging that incorporates cumulative damage over the stress history. The accuracy of the proposed model is evaluated through a comprehensive comparison of numerical results with experimental data. The experimental data set encompasses traditional triaxial compression tests, single-stage creep, multistage creep, and single-stage relaxation tests conducted under varying confining pressures. The predicted results exhibit strong consistency with the entire data set. Furthermore, this paper employs crack damage stress as an indicator characterizing the long-term strength of rock. Through frictional damage coupling analysis and derivation, an analytical expression for the long-term strength of rock materials containing microcracks is provided, serving as a theoretical basis for investigating the long-term mechanical performance of brittle rock materials and ensuring the long-term stability of large-scale rock engineering projects.

Bank Emails: The Language of Legit and Scam

Amidst the pandemic, the surge in unsolicited scam emails has made internet technology a significant concern. Distinguishing between scams and authentic/legit emails has become increasingly difficult. This research aimed to identify typical characteristics and language patterns found in emails. It also aimed to identify the common linguistics and technical features of the scams and legitimate bank emails. Two legitimate emails and two scam or fraudulent emails from different banks in the Philippines (i.e., LandBank and BDO). A qualitative method was used to analyze and interpret the language features of the two emails. This led to the identification of shared characteristics found in scams and legitimate emails. A scam email typically includes embedded malicious links, misspellings, formatting issues, suspicious email domains, frequent use of contractions and redundancy, tautology, and imperative mood. On the other hand, legitimate emails exhibit proper punctuation, accurate email domains, appropriate formatting, absence of spelling mistakes, avoidance of contractions and redundancy, and use indicative mood rather than imperative. Scam emails utilize modal verbs to manipulate recipients into divulging personal information. Recognizing these email patterns is crucial for educating users and reducing phishing risks. It is important to note that the interpretation and findings of this study are based on the entire dataset collected. Keywords: Bank Emails, Legit Emails, Linguistic Features, Scam Emails, Phishing

Detecting outliers in case-control cohorts for improving deep learning networks on Schizophrenia prediction.

This study delves into the intricate genetic and clinical aspects of Schizophrenia, a complex mental disorder with uncertain etiology. Deep Learning (DL) holds promise for analyzing large genomic datasets to uncover new risk factors. However, based on reports of non-negligible misdiagnosis rates for SCZ, case-control cohorts may contain outlying genetic profiles, hindering compelling performances of classification models. The research employed a case-control dataset sourced from the Swedish populace. A gene-annotation-based DL architecture was developed and employed in two stages. First, the model was trained on the entire dataset to highlight differences between cases and controls. Then, samples likely to be misclassified were excluded, and the model was retrained on the refined dataset for performance evaluation. The results indicate that SCZ prevalence and misdiagnosis rates can affect case-control cohorts, potentially compromising future studies reliant on such datasets. However, by detecting and filtering outliers, the study demonstrates the feasibility of adapting DL methodologies to large-scale biological problems, producing results more aligned with existing heritability estimates for SCZ. This approach not only advances the comprehension of the genetic background of SCZ but also opens doors for adapting DL techniques in complex research for precision medicine in mental health.

External validation of a clinical mathematical model estimating post-operative urine output following cardiac surgery in children

BackgroundThis study aims to externally validate a clinical mathematical model designed to predict urine output (UOP) during the initial post-operative period in pediatric patients who underwent cardiac surgery with cardiopulmonary bypass (CPB).MethodsChildren aged 0–18 years admitted to the pediatric cardiac intensive care unit at Cleveland Clinic Children’s from April 2018 to April 2023, who underwent cardiac surgery with CPB were included. Patients were excluded if they had pre-operative kidney failure requiring kidney replacement therapy (KRT), re-operation or extracorporeal membrane oxygenation or KRT requirement within the first 32 post-operative hours or had indwelling urinary catheter for fewer than the initial 32 post-operative hours, or had vasoactive-inotrope score of 0, or those with missing data in the electronic health records.ResultsA total of 213 encounters were analyzed; median age (days): 172 (IQR 25–75th%: 51–1655), weight (kg): 6.1 (IQR 25–75th%: 3.8–15.5), median UOP ml/kg/hr in the first 32 post-operative hours: 2.59 (IQR 25–75th%: 1.93–3.26) and post-operative 30-day mortality: 1, (0.4%). The mathematical model achieved the following metrics in the entire dataset: mean absolute error (95th% Confidence Interval (CI)): 0.70 (0.67–0.73), median absolute error (95th% CI): 0.54 (0.52–0.56), mean squared error (95th% CI): 0.97 (0.89–1.05), root mean squared error (95th% CI): 0.99 (0.95–1.03) and R2 Score (95th% CI): 0.29 (0.24–0.34).ConclusionsThis study provides encouraging external validation results of a mathematical model predicting post-operative UOP in pediatric cardiac surgery patients. Further multicenter studies must explore its broader applicability.Graphical abstractA higher resolution version of the Graphical abstract is available as Supplementary information

Site-Independent Mapping of Clay Content in Vineyard Soils via Mobile Proximal Gamma-Ray Spectrometry and Machine Learning Calibrations.

Vineyards hold considerable soil variability between regions and plots, and there is frequently large soil heterogeneity within plots. Clay content in vineyard soils is of interest with respect to soil management, environmental monitoring, and wine quality. However, spatially resolved clay mapping is laborious and expensive. Gamma-ray spectrometry (GS) is a suitable tool for predicting clay content in precision agriculture when locally calibrated, but it has scarcely been tested site-independently and in vineyards. This study evaluated GS to predict clay content with a site-independent calibration and four machine learning algorithms (Support Vector Machines, Random Forest, k-Nearest Neighbors, and Bayesian regulated neuronal networks) in eight vineyards from four German vine-growing regions. Clay content in the studied soils ranged from 62 to 647 g kg-1. The Random Forest calibration was most suitable. Test set evaluation revealed good model performance for the entire dataset with RPIQ = 4.64, RMSEP = 56.7 g kg-1, and R2 = 0.87; however, prediction quality varied between the sites. Overall, GS with the Random Forest model calibration was appropriate to predict the clay content and its spatial distribution, even for heterogeneous geopedological settings and in individual plots. Therefore, GS is considered a valuable tool for soil mapping in vineyards, where clay content and product quality are closely linked.

Enhancing Interpretability in Medical Image Classification by Integrating Formal Concept Analysis with Convolutional Neural Networks.

In this study, we present a novel approach to enhancing the interpretability of medical image classification by integrating formal concept analysis (FCA) with convolutional neural networks (CNNs). While CNNs are increasingly applied in medical diagnoses, understanding their decision-making remains a challenge. Although visualization techniques like saliency maps offer insights into CNNs' decision-making for individual images, they do not explicitly establish a relationship between the high-level features learned by CNNs and the class labels across entire dataset. To bridge this gap, we leverage the FCA framework as an image classification model, presenting a novel method for understanding the relationship between abstract features and class labels in medical imaging. Building on our previous work, which applied this method to the MNIST handwritten image dataset and demonstrated that the performance is comparable to CNNs, we extend our approach and evaluation to histopathological image datasets, including Warwick-QU and BreakHIS. Our results show that the FCA-based classifier offers comparable accuracy to deep neural classifiers while providing transparency into the classification process, an important factor in clinical decision-making.

ChickenSense: A Low-Cost Deep Learning-Based Solution for Poultry Feed Consumption Monitoring Using Sound Technology

This research proposes a low-cost system consisting of a hardware setup and a deep learning-based model to estimate broiler chickens’ feed intake, utilizing audio signals captured by piezoelectric sensors. The signals were recorded 24/7 for 19 consecutive days. A subset of the raw data was chosen, and events were labeled in two classes, feed-pecking and non-pecking (including singing, anomaly, and silence samples). Next, the labeled data were preprocessed through a noise removal algorithm and a band-pass filter. Then, the spectrogram and the signal envelope were extracted from each signal and fed as inputs to a VGG-16-based convolutional neural network (CNN) with two branches for 1D and 2D feature extraction followed by a binary classification head to classify feed-pecking and non-pecking events. The model achieved 92% accuracy in feed-pecking vs. non-pecking events classification with an f1-score of 91%. Finally, the entire raw dataset was processed utilizing the developed model, and the resulting feed intake estimation was compared with the ground truth data from scale measures. The estimated feed consumption showed an 8 ± 7% mean percent error on daily feed intake estimation with a 71% R2 score and 85% Pearson product moment correlation coefficient (PPMCC) on hourly intake estimation. The results demonstrate that the proposed system estimates broiler feed intake at each feeder and has the potential to be implemented in commercial farms.

Polygenic hazard score models for the prediction of Alzheimer's free survival using the lasso for Cox's proportional hazards model.

The prediction of the susceptibility of an individual to a certain disease is an important and timely research area. An established technique is to estimate the risk of an individual with the help of an integrated risk model, that is, a polygenic risk score with added epidemiological covariates. However, integrated risk models do not capture any time dependence, and may provide a point estimate of the relative risk with respect to a reference population. The aim of this work is twofold. First, we explore and advocate the idea of predicting the time-dependent hazard and survival (defined as disease-free time) of an individual for the onset of a disease. This provides a practitioner with a much more differentiated view of absolute survival as a function of time. Second, to compute the time-dependent risk of an individual, we use published methodology to fit a Cox's proportional hazard model to data from a genetic SNP study of time to Alzheimer's disease (AD) onset, using the lasso to incorporate further epidemiological variables such as sex, APOE (apolipoprotein E, a genetic risk factor for AD) status, 10 leading principal components, and selected genomic loci. We apply the lasso for Cox's proportional hazards to a data set of 6792 AD patients (composed of 4102 cases and 2690 controls) and 87 covariates. We demonstrate that fitting a lasso model for Cox's proportional hazards allows one to obtain more accurate survival curves than with state-of-the-art (likelihood-based) methods. Moreover, the methodology allows one to obtain personalized survival curves for a patient, thus giving a much more differentiated view of the expected progression of a disease than the view offered by integrated risk models. The runtime to compute personalized survival curves is under a minute for the entire data set of AD patients, thus enabling it to handle datasets with 60,000-100,000 subjects in less than 1 h.

Exploring fine-scale urban landscapes using satellite data to predict the distribution of Aedes mosquito breeding sites

BackgroundThe spread of mosquito-transmitted diseases such as dengue is a major public health issue worldwide. The Aedes aegypti mosquito, a primary vector for dengue, thrives in urban environments and breeds mainly in artificial or natural water containers. While the relationship between urban landscapes and potential breeding sites remains poorly understood, such a knowledge could help mitigate the risks associated with these diseases. This study aimed to analyze the relationships between urban landscape characteristics and potential breeding site abundance and type in cities of French Guiana (South America), and to evaluate the potential of such variables to be used in predictive models.MethodsWe use Multifactorial Analysis to explore the relationship between urban landscape characteristics derived from very high resolution satellite imagery, and potential breeding sites recorded from in-situ surveys. We then applied Random Forest models with different sets of urban variables to predict the number of potential breeding sites where entomological data are not available.ResultsLandscape analyses applied to satellite images showed that urban types can be clearly identified using texture indices. The Multiple Factor Analysis helped identify variables related to the distribution of potential breeding sites, such as buildings class area, landscape shape index, building number, and the first component of texture indices. Models predicting the number of potential breeding sites using the entire dataset provided an R² of 0.90, possibly influenced by overfitting, but allowing the prediction over all the study sites. Predictions of potential breeding sites varied highly depending on their type, with better results on breeding sites types commonly found in urban landscapes, such as containers of less than 200 L, large volumes and barrels. The study also outlined the limitation offered by the entomological data, whose sampling was not specifically designed for this study. Model outputs could be used as input to a mosquito dynamics model when no accurate field data are available.ConclusionThis study offers a first use of routinely collected data on potential breeding sites in a research study. It highlights the potential benefits of including satellite-based characterizations of the urban environment to improve vector control strategies.

WebGIS Open-source Platform for Localizations of New P2G Plants in Sicily

The climatic emergency that involves the globe has led targets of greenhouse gas reduction in the EU and all over the world. In this scenario, recent advances in renewable renewable energy sources (RESs) have focused interest on the diffusion of power supplies that are produced by photovoltaic and wind plants. The non-programmable nature of these energy sources has led recent studies to consider the power-to-gas (P2G) solution as an opportunity for employing the curtailed electric energy by converting it into hydrogen. The localizations of P2G plants depend on several factors regarding power production, distances, and population distributions. The necessity of integrating these factors led this work to study the development of a cost function that is hosted by a webbased GIS (geographic information system) platform, thus allowing for the storage, elaboration, and web fruition of an entire data set that is related to the possible new localizations of P2G plants. The structure is based on open-source technology and creates a solution that is easily employable by specialists. The developed platform is composed of different remotely connected blocks that are solely based on open-source technology and is focused the interest on the territory of Sicily (Italy). GIS software, a RDBMS database, a geospatial server (Geoserver ), a Python optimization module, and a WebGIS visualizer are integrated. This work represents a scientific contribution to the management of energy sources, with a particular focus on policies that are based on hydrogen technology. In fact, different data sets that contain several levels of information that are related to the management and the localization of P2G plants will be even further employed in the future.