Earlier Models Research Articles

An efficient model for small object detection in the maritime environment

Environmental perception is crucial for autonomous ships realizing autonomous navigation, in particular, the high-precision and low-latency detection of small objects on the sea surface is a key and challenging issue. To address this problem, this paper presents a model that improves the detection accuracy and delivers excellent real-time performance for autonomous ship navigation. The backbone of the proposed model enhances the modelling capabilities by expanding deformable convolution and introducing the self-designed attention mechanism. Additionally, an enhanced feature fusion structure is designed by the pixel shuffle based on super-resolution reconstruction to keep the integrity of feature information for small objects. This paper also presents an optimized model quantization strategy that alleviates the problem of low model efficiency caused by the limited resources onboard the ship. Compared to the earlier model, the present one has increased the mean average precision on the Rizhao Zhuimeng-3# maritime optical dataset by 4.5 % and by 21 % for small object detection. Furthermore, the real-time detection for high-resolution images can now reach a speed of 67 frames/s. Moreover, the present model outperforms existing methods concerning accuracy when the frames/s are similar. The results indicate that the proposed model can be potentially applied to autonomous ships.

Optimized ensemble deep learning for predictive analysis of student achievement.

Education is essential for individuals to lead fulfilling lives and attain greatness by enhancing their value. It improves self-assurance and enables individuals to navigate the complexities of modern society effectively. Despite the obstacles it faces, education continues to develop. The objective of numerous pedagogical approaches is to enhance academic performance. The development of technology, especially artificial intelligence, has caused a significant change in learning. This has made instructional materials available anytime and wherever easily accessible. Higher education institutions are adding technology to conventional teaching strategies to improve learning. This work presents an innovative approach to student performance prediction in educational settings. The strategy combines the DistilBERT with LSTM (DBTM) hybrid approach with the Spotted Hyena Optimizer (SHO) to change parameters. Regarding accuracy, log loss, and execution time, the model significantly improved over earlier models. The challenges presented by the increasing volume of data in graduate and postgraduate programs are effectively addressed by the proposed method. It produces exceptional performance metrics, including a 15-25% decrease in processing time through optimization, 98.7% accuracy, and 0.03% log loss. This work additionally demonstrates the effectiveness of DBTM-SHO in administering extensive datasets and makes an important improvement to educational data mining. It provides a robust foundation for organizations facing the challenges of evaluating student achievement in the era of vast data.

Cost-consequence analysis of computer vision-based skin prick tests: implications for cost containment in Switzerland

BackgroundSkin prick tests (SPTs), or intraepidermal tests, are often the first diagnostic approach for people with a suspected allergy. Together with the clinical history, SPTs allow doctors to draw conclusions on allergies based on the sensitization pattern. The purpose of this study is to investigate the potential cost consequences that would accrue to a Swiss University hospital after the adoption of computer vision-based SPTs.MethodsWe conducted a cost-consequence analysis from a hospital’s perspective to evaluate the potential cost consequences of using a computer vision-based system to read SPT results. The patient population consisted of individuals who were referred to the allergology department of one of the five university hospitals in Switzerland, Inselspital, whose allergology department averages 100 SPTs a week. We developed an early cost-consequence model comparing two SPT techniques; computer vision-based SPTs conducted with the aid of Nexkin DSPT and standard fully manual SPTs. Probabilistic sensitivity analysis and additional univariate sensitivity analyses were used to account for uncertainty.ResultsThe difference in average cost between the two alternatives from a hospital’s perspective was estimated to be CHF 7 per SPT, in favour of the computer vison-based SPTs. Monte Carlo probabilistic simulation also indicated that SPTs conducted using the computer vision-based system were cost saving compared to standard fully manual SPTs. Sensitivity analyses additionally demonstrated the robustness of the base case result subject to plausible changes in all the input parameters, with parameters representing the costs associated with both SPT techniques having the largest influence on the incremental cost. However, higher sensitization prevalence rates seemed to favour the more accurate standard fully manual SPTs.ConclusionAgainst the backdrop of rising healthcare costs especially in Switzerland, using computer-aided or (semi) automated diagnostic systems could play an important role in healthcare cost containment efforts. However, results should be taken with caution because of the uncertainty associated with the early nature of our analysis and the specific Swiss context adopted in this study.

High-redshift Merger Model for Low-frequency Gravitational Wave Background

In 2023, the Pulsar Timing Array Collaborations announced the discovery of a gravitational wave background (GWB), predominantly attributed to supermassive black hole binary (SMBHB) mergers. However, the detected GWB is several times stronger than the default value expected from galactic observations at low and moderate redshifts. Recent findings by the James Webb Space Telescope have unveiled a substantial number of massive, high-redshift galaxies, suggesting more massive SMBHB mergers at these early epochs. Motivated by these findings, we propose an “early merger” model that complements the standard merger statistics by incorporating these early, massive galaxies. We compare the early and standard “late merger” models, which assume peak merger rates in the local Universe, and match both merger models to the currently detected GWB. Our analysis shows that the early merger model has a significantly lower detection probability for single binaries and predicts a ∼30% likelihood that the first detectable single source will be highly redshifted and remarkably massive with rapid frequency evolution. In contrast, the late merger model predicts a nearly monochromatic first source at low redshift. The future confirmation of an enhanced population of massive high-redshift galaxies and the detection of fast-evolving binaries would strongly support the early merger model, offering significant insights into the evolution of galaxies and SMBHs.

Memory-Based Dynamic Bayesian Networks for Learner Modeling: Towards Early Prediction of Learners’ Performance in Computational Thinking

Artificial intelligence (AI) has demonstrated significant potential in addressing educational challenges in digital learning. Despite this potential, there are still concerns about the interpretability and trustworthiness of AI methods. Dynamic Bayesian networks (DBNs) not only provide interpretability and the ability to integrate data-driven insights with expert judgment for enhanced trustworthiness but also effectively process temporal dynamics and relationships in data, crucial for early predictive modeling tasks. This research introduces an approach for the temporal modeling of learners’ computational thinking abilities that incorporates higher-order influences of latent variables (hereafter referred to as memory of the model) and accordingly predicts learners’ performance early. Our findings on educational data from the AutoThinking game indicate that when using only first-order influences, our proposed model can predict learners’ performance early, with an 86% overall accuracy (i.e., time stamps 0, 5, and 9) and a 94% AUC (at the last time stamp) during cross-validation and 91% accuracy and 98% AUC (at the last time stamp) in a holdout test. The introduction of higher-order influences improves model accuracy in both cross-validation and holdout tests by roughly 4% and improves the AUC at timestamp 0 by roughly 2%. This suggests that integrating higher-order influences into a DBN not only potentially improves the model’s predictive accuracy during the cross-validation phase but also enhances its overall and time stamp-specific generalizability. DBNs with higher-order influences offer a trustworthy and interpretable tool for educators to foresee and support learning progression.

A Review of Early Gas Intrusion Detection based on Multiphase Flow Model

High content gas reservoirs in deep reservoirs have complex pressure systems, and gas penetration occurs frequently during drilling, trip or other drilling operations. Due to the compressibility of gas and its easy solubility in drilling fluid, gas penetration in ultra-deep Wells is often sudden and hidden. In the early stage, it is difficult to find and fully alleviate gas invasion, which is very easy to induce blowout accidents. Accurate prediction and control of gas invasion is of great significance to prevent blowout accidents. In the past 20 years, the research of early gas invasion detection has made great progress. This paper aims to provide the latest progress of early gas intrusion detection based on transient multiphase flow, and make a comprehensive review of early gas intrusion detection. The specific contents include: 1. 2. The development of numerical simulation of early gas invasion detection is summarized. 3. The influencing factors of early gas intrusion detection modeling are analyzed. 4. The application of drift flux model in each flow pattern is introduced. 5. The importance of gas solubility and heat transfer in numerical simulation of gas penetration detection is analyzed. 6. The current development of early gas invasion detection was summarized and the future prospect was made.

Enhancing Biomedical Question Answering with Large Language Models

In the field of Information Retrieval, biomedical question answering is a specialized task that focuses on answering questions related to medical and healthcare domains. The goal is to provide accurate and relevant answers to the posed queries related to medical conditions, treatments, procedures, medications, and other healthcare-related topics. Well-designed models should efficiently retrieve relevant passages. Early retrieval models can quickly retrieve passages but often with low precision. In contrast, recently developed Large Language Models can retrieve documents with high precision but at a slower pace. To tackle this issue, we propose a two-stage retrieval approach that initially utilizes BM25 for a preliminary search to identify potential candidate documents; subsequently, a Large Language Model is fine-tuned to evaluate the relevance of query–document pairs. Experimental results indicate that our approach achieves comparative performances on the BioASQ and the TREC-COVID datasets.

Regulation of glycolysis-derived L-lactate production in astrocytes rescues the memory deficits and Aβ burden in early Alzheimer’s disease models

Aberrant energy metabolism in the brain is a common pathological feature in the preclinical Alzheimer's Disease (AD). Recent studies have reported the early elevations of glycolysis-involved enzymes in AD brain and cerebrospinal fluid according to a large-scale proteomic analysis. It’s well-known that astrocytes exhibit strong glycolytic metabolic ability and play a key role in the regulation of brain homeostasis. However, its relationship with glycolytic changes and cognitive deficits in early AD patients is unclear. Here, we investigated the mechanisms by which astrocyte glycolysis is involved in early AD and its potential as a therapeutic target. Our results suggest that Aβ-activated microglia can induce glycolytic-enhanced astrocytes in vitro, and that these processes are dependent on the activation of the AKT-mTOR-HIF-1α pathway. In early AD models, the increase in L-lactate produced by enhanced glycolysis of astrocytes leads to spatial cognitive impairment by disrupting synaptic plasticity and accelerating Aβ aggregation. Furthermore, we find rapamycin, the mTOR inhibitor, can rescue the impaired spatial memory and Aβ burden by inhibiting the glycolysis-derived L-lactate in the early AD models. In conclusion, we highlight that astrocytic glycolysis plays a critical role in the early onset of AD and that the modulation of glycolysis-derived L-lactate by rapamycin provides a new strategy for the treatment of AD.

Early sepsis mortality prediction model based on interpretable machine learning approach: development and validation study.

Sepsis triggers a harmful immune response due to infection, causing high mortality. Predicting sepsis outcomes early is vital. Despite machine learning's (ML) use in medical research, local validation within the Medical Information Mart for Intensive Care IV (MIMIC-IV) database is lacking. We aimed to devise a prognostic model, leveraging MIMIC-IV data, to predict sepsis mortality and validate it in a Chinese teaching hospital. MIMIC-IV provided patient data, split into training and internal validation sets. Four ML models logistic regression (LR), support vector machine (SVM), deep neural networks (DNN), and extreme gradient boosting (XGBoost) were employed. Shapley additive interpretation offered early and interpretable mortality predictions. Area under the ROC curve (AUROC) gaged predictive performance. Results were cross verified in a Chinese teaching hospital. The study included 27,134 sepsis patients from MIMIC-IV and 487 from China. After comparing, 52 clinical indicators were selected for ML model development. All models exhibited excellent discriminative ability. XGBoost surpassed others, with AUROC of 0.873 internally and 0.844 externally. XGBoost outperformed other ML models (LR: 0.829; SVM: 0.830; DNN: 0.837) and clinical scores (Simplified Acute Physiology Score II: 0.728; Sequential Organ Failure Assessment: 0.728; Oxford Acute Severity of Illness Score: 0.738; Glasgow Coma Scale: 0.691). XGBoost's hospital mortality prediction achieved AUROC 0.873, sensitivity 0.818, accuracy 0.777, specificity 0.768, and F1 score 0.551. We crafted an interpretable model for sepsis death risk prediction. ML algorithms surpassed traditional scores for sepsis mortality forecast. Validation in a Chinese teaching hospital echoed these findings.

A 1D-CNN model for the early detection of citrus Huanglongbing disease in the sieve plate of phloem tissue using micro-FTIR

Among the most frequently diagnosed diseases in citrus, citrus Huanglongbing disease has caused severe economic losses to the citrus industry worldwide since there is no curable method and it spreads quickly. As callose accumulation in phloem is one of the early response events to Asian species Candidatus Liberibacter asiaticus (CLas) infection, the dynamic perception of the sieve plate region can be used as an indicator for the early diagnosis of citrus HLB disease. In this study, one-dimensional convolutional neural network (1D-CNN) models were established to achieve early detection of HLB disease based on spectral information in the sieve plate region using Fourier transform infrared microscopy (micro-FTIR) spectrometer. Partial least squares regression (PLSR) and the least squares support vector machine regression (LS-SVR) models are used for the prediction of callose based on the micro-FTIR information in the sieve plate region of the citrus midrib. Furthermore, an improved data augmentation method by superimposing Gaussian noise was proposed to expand the spectral amplitude. The proposed method has achieved 98.65 % classification accuracy, which was higher than that of other traditional algorithms such as the logistic model tree (LMT), linear discriminant analysis (LDA), Bayes (BS), support vector machine (SVM) and k-nearest neighbors (kNN), and also than that of the molecular detection qPCR (Quantitative real-time polymerase chain reaction) method. Finally, based on the established early detection model with laboratory samples, it can also be used to detect the citrus HLB in complex field samples by using model updating methods, and the overall detection accuracy of the model reached 91.21 %. Our approach has potential for the early diagnosis of citrus HLB disease from the microscopic scale, which would provide useful and precise guidelines to prevent and control citrus HLB disease.

Plasma lipidomics in early APP/PS1 female mouse model and its relationship with brain: Is it affected by the estrous cycle?

BackgroundAlzheimer's disease (AD) is the most prevalent dementia, showing higher incidence in women. Besides, lipids play an essential role in brain, and they could be dysregulated in neurodegeneration. Specifically, impaired plasma lipid levels could predict early AD diagnosis. This work aims to identify the main plasma lipids altered in early AD female mouse model and evaluate their relationship with brain lipidome. Also, the possible involvement of the estrous cycle in lipid metabolism has been evaluated.MethodsPlasma samples of wild-type (n = 10) and APP/PS1 (n = 10) female mice of 5 months of age were collected, processed, and analysed using a lipidomic mass spectrometry-based method. A statistical analysis involving univariate and multivariate approaches was performed to identify significant lipid differences related to AD between groups. Also, cytology tests were conducted to confirm estrous cycle phases.ResultsThree hundred thirty lipids were detected in plasma, 18 of them showed significant differences between groups; specifically, some triacylglycerols, cholesteryl esters, lysophosphatidylcholines, phosphatidylcholines, and ether-linked phosphatidylcholines, increased in early AD; while other phosphatidylcholines, phosphatidylethanolamines, ceramides, and ether-linked phosphatidylethanolamines decreased in early AD. A multivariate approach was developed from some lipid variables, showing high diagnostic indexes (70% sensitivity, 90% specificity, 80% accuracy). From brain and plasma lipidome, some significant correlations were observed, mainly in the glycerophospholipid family. Also, some differences were found in both plasma and brain lipids, according to the estrous cycle phase.ConclusionsTherefore, lipid alterations can be identified in plasma at early AD stages in mice females, with a relationship with brain lipid metabolism for most of the lipid subfamilies, suggesting some lipids as potential AD biomarkers. In addition, the estrous cycle monitoring could be relevant in female studies.

FolA thyA knockout E. coli as a suitable surrogate model for evaluation of antifolate sensitivity against PfDHFR-TS

A new superior bacteria complementation model was achieved for testing antifolate compounds and investigating antifolate resistance in the dihydrofolate reductase (DHFR) enzyme of the malaria parasite. Earlier models depended on the addition of trimethoprim (TMP) to chemically suppress the host Escherichia coli (Ec) DHFR function. However, incomplete suppression of EcDHFR and potential interference of antibiotics needed to maintain plasmids for complementary gene expression can complicate the interpretations. To overcome such limitations, the folA (F) and thyA (T) genes were genetically knocked out (Δ) in E. coli BL21(DE3). The resulting EcΔFΔT cells were thymidine auxotroph where thymidine supplementation or functional complementation with heterologous DHFR-thymidylate synthase (TS) is needed to restore the loss of gene functions. When tested against pyrimethamine (PYR) and its analogs designed to target Plasmodium falciparum (Pf) DHFR-TS, the 50 % inhibitory concentration values obtained from EcΔFΔT surrogates expressing wildtype (PfTM4) or double mutant (PfK1) DHFR-TS showed strong correlations to the results obtained from the standard in vitro P. falciparum growth inhibition assay. Interestingly, while TMP had little effect on the susceptibility to PYR and analogs in EcΔFΔT expressing PfDHFR-TS, it hypersensitized the chemically knockdown E. coli BL21(DE3) expressing PfTM4 DHFR-TS but desensitized the one carrying PfK1 DHFR-TS. The low intrinsic expression level of PfTM4 in E. coli BL21(DE3) by western blot analysis may explain the hypersensitive to antifolates of chemical knockdown bacteria surrogate. These results demonstrated the usefulness of EcΔFΔT surrogate as a new tool for antifolate antimalarial screening with potential application for investigation of antifolate resistance mechanism.

Predicting the translation efficiency of messenger RNA in mammalian cells.

The degree to which translational control is specified by mRNA sequence is poorly understood in mammalian cells. Here, we constructed and leveraged a compendium of 3,819 ribosomal profiling datasets, distilling them into a transcriptome-wide atlas of translation efficiency (TE) measurements encompassing >140 human and mouse cell types. We subsequently developed RiboNN, a multitask deep convolutional neural network, and classic machine learning models to predict TEs in hundreds of cell types from sequence-encoded mRNA features, achieving state-of-the-art performance (r=0.79 in human and r=0.78 in mouse for mean TE across cell types). While the majority of earlier models solely considered 5' UTR sequence, RiboNN integrates contributions from the full-length mRNA sequence, learning that the 5' UTR, CDS, and 3' UTR respectively possess ~67%, 31%, and 2% per-nucleotide information density in the specification of mammalian TEs. Interpretation of RiboNN revealed that the spatial positioning of low-level di- and tri-nucleotide features (i.e., including codons) largely explain model performance, capturing mechanistic principles such as how ribosomal processivity and tRNA abundance control translational output. RiboNN is predictive of the translational behavior of base-modified therapeutic RNA, and can explain evolutionary selection pressures in human 5' UTRs. Finally, it detects a common language governing mRNA regulatory control and highlights the interconnectedness of mRNA translation, stability, and localization in mammalian organisms.

Procedural containment vs. substantive entrenchment: two early models of militant democracy

ABSTRACT German lawyer and political scientist Karl Loewenstein is generally regarded as the originator of the militant democracy paradigm. In a series of articles in the mid and late 1930s, he argued that constitutional democracies should pre-emptively defend themselves against movements and parties that were seeking to undermine them. More recently, another father of the paradigm has been identified in the controversial figure of Carl Schmitt. Before his despicable and opportunistic support for the Nazi regime at the end of 1933, he advocated a model of authoritarian democracy based on fundamental constitutional contents that should be made impervious to legislative change. This article examines the different models of disciplined and authoritarian democracy proposed by these authors. While both advocated measures that could inhibit the activity of parliamentary majorities composed mainly of parties opposed to the state constitutional structure, Schmitt, more than Loewenstein, made the point that the defence of democracy requires subscribing to substantive values and principles and entrenching them. The juxtaposition of these models makes it clear that militant democracy not only recommends the protection of certain procedures but also entails a normative standard comprising substantive contents. This helps to highlight the unstated normative burdens of democratic militancy.

Reactant Discovery with an Ab Initio Nanoreactor: Exploration of Astrophysical N-Heterocycle Precursors and Formation Pathways.

The incorporation of nitrogen atoms into cyclic compounds is essential for terrestrial life; nitrogen-containing (N-)heterocycles make up DNA and RNA nucleobases, several amino acids, B vitamins, porphyrins, and other components of biomolecules. The discovery of these molecules on meteorites with non-terrestrial isotopic abundances supports the hypothesis of exogenous delivery of prebiotic material to early Earth; however, there has been no detection of these species in interstellar environments, indicating that there is a need for greater knowledge of their astrochemical formation and destruction pathways. Here, we present results of simulations of gas-phase pyrrole and pyridine formation from an ab initio nanoreactor, a first-principles molecular dynamics simulation method that accelerates reaction discovery by applying non-equilibrium forces that are agnostic to individual reaction coordinates. Using the nanoreactor in a retrosynthetic mode, starting with the N-heterocycle of interest and a radical leaving group, then considering the discovered reaction pathways in reverse, a rich landscape of N-heterocycle-forming reactivity can be found. Several of these reaction pathways, when mapped to their corresponding minimum energy paths, correspond to novel barrierless formation pathways for pyridine and pyrrole, starting from both detected and hypothesized astrochemical precursors. This study demonstrates how first-principles reaction discovery can build mechanistic knowledge in astrochemical environments as well as in early Earth models such as Titan's atmosphere where N-heterocycles have been tentatively detected.

Multimodal Religiously Hateful Social Media Memes Classification Based on Textual and Image Data

Multimodal hateful social media meme detection is an important and challenging problem in the vision-language domain. Recent studies show high accuracy for such multimodal tasks due to datasets that provide better joint multimodal embedding to narrow the semantic gap. Religiously hateful meme detection is not extensively explored among published datasets. While there is a need for higher accuracy on religiously hateful memes, deep learning–based models often suffer from inductive bias. This issue is addressed in this work with the following contributions. First, a religiously hateful memes dataset is created and published publicly to advance hateful religious memes detection research. Over 2000 meme images are collected with their corresponding text. The proposed approach compares and fine-tunes VisualBERT pre-trained on the Conceptual Caption (CC) dataset for the downstream classification task. We also extend the dataset with the Facebook hateful memes dataset. We extract visual features using ResNeXT-152 Aggregated Residual Transformations–based Masked Regions with Convolutional Neural Networks (R-CNN) and Bidirectional Encoder Representations from Transformers (BERT) uncased for textual encoding for the early fusion model. We use the primary evaluation metric of an Area Under the Operator Characters Curve (AUROC) to measure model separability. Results show that the proposed approach has a higher AUROC score of 78%, proving the model’s higher separability performance and an accuracy of 70%. It shows comparatively superior performance considering dataset size and against ensemble-based machine learning approaches.

Self-consistent theory of cosmic ray penetration into molecular clouds: Relativistic case

We study penetration of interstellar cosmic rays (CRs) into molecular clouds surrounded by nonuniform diffuse envelopes. The present work generalizes our earlier model of CR self-modulation [Ivlev ; Dogiel ], in which the value for the envelope’s gas density where CRs excite MHD waves was treated as a free parameter. Now, we investigate the case where the density monotonically increases toward the center. Assuming that CRs are relativistic, we obtain a universal analytical solution which does not depend on the particular shape of gas distribution in the envelope, and self-consistently derive boundaries of the diffusion zone formed within the envelope, where CRs are scattered at the self-excited waves. The values of the gas density at the boundaries are found to be substantially smaller than those assumed in the earlier model, which leads to a significantly stronger modulation of penetrating CRs. We compute the impact of CR self-modulation on the gamma-ray emission and show that the results of our theoretical model are in excellent agreement with recent observations of nearby giant molecular clouds by Yang []. Published by the American Physical Society 2024

Community-Based Early Warning System Model for Stream Overflow In Barranquilla

Context: This work aims to design and create a community-based early warning model as an alternative for the mitigation of disasters caused by stream overflow in Barranquilla (Colombia). This model is based on contributions from social networks, which are consulted through their API and filtered according to their location. Methods: With the information collected, cleaning and debugging are performed. Then, through natural language processing techniques, the texts are tokenized and vectorized, aiming to find the vector similarity between the processed texts and thus generating a classification. Results: The texts classified as dealing with stream overflow are processed again to obtain a location or assign a default one, in order to for them to be georeferenced in a map that allows associating the risk zone and visualizing it in a web application to monitor and reduce the potential damage to the population. Conclusions: Three classification algorithms were selected (random forest, extra trees, and k-neighbors) to determine the best classifier. These three algorithms exhibited the best performance and R2 regarding the data processed in the regressions. These algorithms were trained, with the k-neighbor algorithm exhibiting the best performance.

Optimal early retirement with target wealth

An agent often considers retirement adequacy when making a retirement decision. This paper introduces a target wealth constraint, where financial wealth must exceed an agent-based threshold for retirement, and studies several early retirement models under such a constraint. We present explicit characterizations of the optimal retirement time and demonstrate the influence of the target wealth on the early retirement decision and the consumption-investment strategy. As the target wealth for retirement increases, the agent will increase consumption, decrease investment in the risky asset, and delay the retirement time. Due to the early retirement effect, the proportion of financial wealth invested in risky assets could increase when the financial wealth approaches the target wealth. Our models demonstrate that retirement adequacy is a crucial factor in early retirement decisions.

An early sepsis prediction model utilizing machine learning and unbalanced data processing in a clinical context

BackgroundEarly and accurate diagnoses of sepsis patients are essential to reduce the mortality. However, the sepsis is still diagnosed in a traditional way in China despite the increasing number of related studies, which may to some extent lead to delays in the treatment. MethodsThe study included 2,385 patients, including 364 with sepsis, collected from the First Affiliated Hospital of Anhui Medical University and partner hospitals from April to July 2022. External validation was conducted using the MIMIC-III database (over 60,000 patients from 2001 to 2012) and the eICU Collaborative Research Database (139,000 patients from 2014 to 2015). Multiple algorithm models, along with the SHapley Additive exPlanations (SHAP) analysis, are applied to explore the main risk factors for the accurate prediction of the sepsis. Multiple Imputations for filling missing data and the Synthetic Minority Oversampling (SMOTE) balancing method for balancing data are used for the data processing. ResultEighteen diagnostic features are used in the predictive model for early sepsis. The Random Forest model has the best performance among all the models, with an Area Under the Curve (AUC) of 87% and an F1-score (F1) of 77%. Moreover, the interpretation from the SHAP analysis is generally consistent with the current clinical situation. ConclusionThe study revealed the relationship between these 18 clinical features and diagnostic outcomes. The results indicate that patients with laboratory values of Systolic Blood Pressure, Albumin, and Heart Rate exceeding certain thresholds are at a high likelihood of developing sepsis.