Integrated Model Research Articles

Artificial intelligence aided serum protein electrophoresis analysis of Finnish patient samples: Retrospective validation.

Serum protein electrophoresis interpretation requires a substantial amount of manual work. In 2020, Chabrun et al. created a machine learning method called SPECTR for the task. We aimed to validate and test the SPECTR method against our results of more precise immunofixation electrophoresis. We gathered 34 625 patients and their first serum protein electrophoresis sample in Helsinki University Hospital. We trained three neural network models: (1) a fractionation model to fractionate electropherograms; (2) a classification model to classify samples to normal, ambiguous, and abnormal (i.e. containing paraprotein); (3) an integration model to predict concentration and location of paraproteins. The fractionation model demonstrated an error rate of ≤0.33g/L in 95% samples. The classification model achieved an area under the curve of 97% in receiver operating characteristic analysis. The integration model demonstrated a coefficient of determination (R2) of 0.991 and a root-mean-square error of 1.37g/L in linear regression. The neural network models proved to be suitable for partial automation in serum protein electrophoresis reporting, i.e. classification of normal electropherograms. Furthermore, the models can accurately suggest the location and concentration of paraproteins.

Traffic Signal and Autonomous Vehicle Control Model: An Integrated Control Model for Connected Autonomous Vehicles at Traffic-Conflicting Intersections Based on Deep Reinforcement Learning

Traffic Signal and Autonomous Vehicle Control Model: An Integrated Control Model for Connected Autonomous Vehicles at Traffic-Conflicting Intersections Based on Deep Reinforcement Learning

Evaluating the cost, cost-effectiveness and survival of an eHealth-facilitated integrated care model for allogeneic stem cell transplantation: Results of the German SMILe randomized, controlled implementation science trial.

eHealth-facilitated integrated care models (eICMs) have proved effective in improving outcomes for chronically ill patients. However, evidence on cost-effectiveness of eICMs is scarce so far. Allogeneic stem cell transplantation (alloSCT) recipients' post-discharge treatment costs and mortality are greatly influenced by complications. Within the international, multicentric SMILe implementation science project, the eHealth-facilitated SMILe integrated care model (SMILe-ICM) was developed to support patients minimize complications' effects within the first year post-alloSCT. Using initial effectiveness findings from the first center that implemented the SMILe-ICM, this study provides a cost and cost-effectiveness evaluation considering one-year and long-term survival effects, post-discharge costs, and patient-related factors. A single-center hybrid effectiveness implementation randomized controlled trial was conducted at a German university hospital from 2/2020 to 8/2022. Eligible alloSCT patients were randomized to the SMILe-ICM or usual care, i.e., one pre-transplant educational nursing visit followed by a physician-led follow-up. The intervention group received usual care plus the SMILe-ICM's four intervention modules (i.e., monitoring of medical/symptom-related parameters, medication adherence, infection prevention, physical activity). All modules were delivered by Advanced Practice Nurses (APNs) in face-to-face visits, combined with continuous online support. Daily, patients entered seventeen medical and symptom-related parameters to the SMILe App, so that APNs could monitor for and investigate possible pre-complication signs. Healthcare utilization costs were assessed at eight time-points (d+30 post-alloSCT-d365) on fourteen self-reported cost indicators and validated against health records. To calculate costs, we applied German standardized unit costs. Cost- and cost-effectiveness were analyzed in five steps: 1.) Calculate total costs, including for the alloSCT inpatient stay and post-discharge follow-up. 2.) Determine life-years gained (survival) as a health benefit unit. 3.) Calculate overall and rehospitalization-free survival estimates. 4.) Calculate the intervention's long-term cost-effectiveness, including extended follow-up, rate of survival until day 1000, and restricted mean survival time. 5.) Contrast these long-term estimates to current post-discharge costs with comparable patient-related factors (age≥or<65, living alone, gender). Seventy-two patients participated (n=36/group). Total intergroup healthcare utilization and post-discharge costs differed, but non-significantly. Survival rates improved with the SMILe-ICM (88% vs. 80%) at least until day +1000. Rehospitalization-free survival showed improvement (38% vs. 30%); however, considering this sample size, both findings were nonsignificant. Cost-effectiveness analysis showed an overall post-discharge cost-effectiveness of 35,364.01€/patient and 6,742€/life year gained - a mean of 79.21 additional days of life for an intervention investment of 1.464€/patient in the first year post-alloSCT. One-year cost-effectiveness was highest for patients living alone. Younger age correlated with longer survival but higher costs. The SMILe-ICM appears to offer survival and rehospitalization benefits, particularly for vulnerable groups, e.g., patients living alone. Larger, adequately powered studies are needed to validate these findings.

Single-task regression naturally adapts to multi-species (eco)toxicological modelling: a case study on animals.

In silico (eco)toxicological modelling has gained increasing popularity with chemical environmentalists in accelerating toxicity assessment of hazardous chemicals on environments, animal well-being and human health. Existing local and multi-task models commonly exhibit restricted extensibility in multi-species modelling scenarios. In this work, we propose a strategy of single-task regression to naturally adapt modelling to (eco)toxicological measurements on multiple species without requiring a certain number of common pesticides among tested species as multi-task regression does. This strategy treats all species equally in an integral model to facilitate data augmentation and inter-species transfer of common patterns of fragmental toxicities. We aggregate 37,305 measurements of 29,140 pesticides on 10 tested groups of animals to train four machine learning models including extreme gradient boosting (XGBoost), deep neural networks (DNN), random forest (RF) and support vector regression (SVR). Five-fold stratified cross-validation shows that the XGBoost outperforms the other three models with overall 0.67 R2, 0.44 RMSE and 0.29 MAE. As compared to local models focusing on one animal group, the proposed single-task regression model achieves a 0.08 ~ 0.49 R2 increase. XGBoost feature importance shows that Morgan bit 389 (five-atom fraction of the aromatic ring) exhibits top importance to single-task regression and single-animal regression. Lastly, taking the pesticide parathion and dimethoate as control baselines, we demonstrate the credibility of several case studies from the viewpoints of toxicity profile similarities and pesticide structural similarities.

A predictive krill distribution model for Euphausia pacifica and Thysanoessa spinifera using scaled acoustic backscatter in the Northern California Current

Euphausiids (krill) are globally significant zooplankton prey for many commercially important or endangered predator species. In the productive upwelling system of the Northern California Current (NCC), two krill species, Euphausia pacifica and Thysanoessa spinifera, dominate the preyscape and constitute an important food resource for many seabirds, cetaceans, and fish. In this study, we use five years of hydroacoustic and net tow data collected in the NCC to develop integrative models predicting acoustic backscatter scaled for E. pacifica or T. spinifera separately. Boosted Regression Trees and Generalized Additive Models are applied in an original ensemble hurdle framework to predict krill presence and abundance from a diverse set of topographic and oceanographic predictors. Krill metrics had significant relationships with seabed depth, distance to submarine canyons, and variables indicative of dynamic ocean conditions (e.g., total deviance explained in acoustic data: 25 % in the presence-absence model & 49 % in the abundance model). Predictions of krill abundance at 5 km resolution averaged by month indicate differential habitat preferences between the two species: T. spinifera was constrained to the continental shelf, around and inshore of the 200 m isobath, whereas E. pacifica was found in greater abundances just offshore of the 200 m isobath and into offshore water in lower abundances. E. pacifica was generally more abundant than T. spinifera (10:1.3 ratio). Both species increased in abundance in the spring and summer, followed by a rapid decline in the fall, and lowest abundances in the winter. These models can produce fine-scale spatial and year-round weekly predictions of E. pacifica and T. spinifera abundance in the NCC, which will provide essential knowledge and new spatial layers about critical ecosystem components to support research and management.

Systematic Review: Integrated Models of Care for Managing Irritable Bowel Syndrome.

Multidisciplinary integrated models of care show promise for improving symptoms and quality of life (QoL) in adults with irritable bowel syndrome (IBS). To describe and evaluate the characteristics of integrated models of care for IBS and identify how digital health is being used in these models of care. Four databases were searched to March 2024 for studies that included adults with IBS who participated in multidisciplinary integrated models of care that delivered non-pharmacological therapies. The template for intervention description and replication (TIDieR) checklist was used to appraise study quality and extract model of care characteristics, which were mapped against the Project INTEGRATE framework to establish topics. Sixteen studies (6 randomized controlled trials, 2 quasi-experimental, 8 cohort studies) reported 14 integrated models of care including 2165 patients of which 918 were IBS patients. Integrated models of care led to improved IBS symptoms (n = 11/13 models of care) and QoL (n = 6/9 models of care). Studies showed moderate compliance with the TIDieR checklist. Five topics were established: clinicians involved, therapies provided, location and mode of delivery, coordinating clinical partnerships, and sharing visions and values of integrated care. Most commonly, a gastroenterologist coordinated care with a psychologist, dietitian, and/or nurse in tertiary care. Psychological, dietary, and physical therapies were provided by n = 11, n = 8, and n = 3 integrated models of care, respectively. Six models of care provided joint consultations or group sessions. Four models of care used digital health such as telephone coaching or online modules. Integrated models of care for IBS exhibited diverse characteristics including the clinicians involved, the therapies provided and the mode of delivery of each therapy. There is a need to evaluate the use of digital health and the delivery of integrated models of care in primary care settings.

What motivates academics in Egypt toward generative AI tools? An integrated model of TAM, SCT, UTAUT2, perceived ethics, and academic integrity

What motivates academics in Egypt toward generative AI tools? An integrated model of TAM, SCT, UTAUT2, perceived ethics, and academic integrity

Rumination Mediates the Relation of Hostile Attribution to Psychological Maladjustment Among Adolescents from Three Countries.

Addressing global concerns about youth mental health requires understanding longitudinal pathways to psychological maladjustment among diverse youth. Hostile attribution bias (HAB) and hostile rumination (HR) are cognitive vulnerabilities associated with multiple forms of psychological maladjustment among diverse youth. This study longitudinally examined whether HR mediates the relation of HAB to aggression, anxiety, and depression symptoms in a sample of adolescents from three countries. Participants included 532 mothers, 384 fathers, and 566 youth (50% female) from Colombia, Italy (Naples and Rome), and the U.S. (White, Black, and Latinx). Structural equation modeling indicated that youth-reported HR at Time 2 (mean age = 12.58 years) significantly mediated the relation of youth-reported HAB at Time 1 (mean age = 10.89 years) to parent-reported aggression and anxiety symptoms at Time 3 (mean age = 13.71 years; aggression: b = 0.05, 95% bootstrap CI = [0.006, 0.14]; anxiety: b = 0.06, 95% bootstrap CI = [0.01, 0.16]); but not to parent-reported depression symptoms at Time 3 (b = 0.02, 95% bootstrap CI = [-0.04, 0.08]). A reverse model indicated HAB at Time 3 significantly mediated the relation of HR at Time 2 to anxiety symptoms at Time 4 (mean age = 14.99 years; b = -0.01, 95% bootstrap CI = [-0.04, -0.001]), but not to aggression or depression symptoms at Time 4. Multi-group analyses indicated focal mediational paths did not vary significantly across national, regional, and racial subgroups or gender. Findings support the Integrative Cognitive Model of Aggression, providing evidence that HAB and HR may be vulnerability factors for aggression and anxiety among diverse youth.

Integrated site selection model for industrial areas: case study for İnegöl furniture industry.

Industrial activities in the central area have adverse effects such as noise, odor, and traffic congestion. Simultaneously, due to changing technological and economic advances, existing industrial areas cannot meet the needs, spatial inadequacies obstruct competition, and production capacity decreases. Decentralizing industrial activities from urban centers are ecologically and economically necessary. Various elements on a macro and micro scale need to be considered to select suitable sites for new industrial areas. Natural, socioeconomic, and built environment features must be examined to ensure sustainability. The objective of this study is to develop an integrated industrial site location model that considers the needs of authorities and industrial stakeholders, as well as economic and ecological sustainability for the İnegöl district, one of Turkey's leading settlements in the furniture industry. Thirty-seven criteria were evaluated using GIS based multi-criteria decision making methods. The criteria were defined through spatial analysis, expert opinions, and in-depth interviews with industry and local government representatives. Using weighted linear combination process the five sub-regions exhibiting the lowest economic costs and the least environmental degradation have been identified. Advantages and disadvantages were identified through the use of sketches and comparisons between the sub-regions. A decision support system was developed for local and central government institutions to be used in industrial site selection processes.

Canonical integral models for Shimura varieties of toral type

Canonical integral models for Shimura varieties of toral type

Innovative education policy models for migrant integration: Bridging access, equity and multicultural inclusion in host country education systems

Innovative education policy models for migrant integration: Bridging access, equity and multicultural inclusion in host country education systems

Understanding the complexity of centers for teaching and learning: introducing a four-dimensional model

Abstract Centers for Teaching and Learning (CTLs) are gaining visibility in the higher education landscape as key actors that support universities in providing high-quality education and fostering a culture that values teaching and learning. While existing research on CTLs typically focuses on evaluation, impact, or institutionally context-specific work, we sought to unearth a framework that could be universally applied to understanding CTL work and functionality. Following data collection and 25 interviews with CTL Directors from higher education institutions across the USA, we created an integrative approach that brings together organizational, behavioral, and relational perspectives. The four dimensions—governance, identity, social capital, and activities—enable us to investigate CTLs from various angles and at different levels, providing valuable clues for understanding CTL performance and challenges. This integrative model can also be used as a framework to inform CTL design decisions.

P-2189. The Integration of Hepatitis C Treatment into Rural Syringe Services Programs via Telehealth: A Highly Effective Care Model

Abstract Background Direct acting antivirals (DAA) effectively cure Hepatitis C (HCV) infection; however, uptake among people who inject drugs (PWID) is low due to social determinants of health and healthcare system barriers. We present a telehealth model of HCV care integration into syringe services programs (SSPs).Figure 1. The HCV Care Continuum through ACCESS Telehealth, a program to integrate hepatitis C and opioid use disorder care into rural Maryland syringe services programs.Note: At time of analysis 6 were pending insurance approval, 15 patients were still on treatment, and 10 had not yet reached the time point for SVR 12 assessment. Methods The Johns Hopkins Viral Hepatitis Center implemented ACCESS Telehealth in collaboration with the Maryland Department of Health Center for Harm Reduction Services. Partnering with individual SSPs, context evaluation, HCV testing workflows, referral pathways for HCV/Opioid Use Disorder (OUD) care, SSP staff training, and SSP-based private space for telemedicine-equipped encounters were established. Services began in August 2021 in three rural Maryland SSPs and expanded to two more in 2022. SSP staff refer clients to low threshold, HCV/OUD treatment. An off-site nurse case manager provides patient education and support, collaborating with SSP-based Certified Peer Recovery Specialists. Specialists offer HCV/OUD treatment at the SSPs via telemedicine. Appointment, prescription, and laboratory information were extracted from the electronic medical record. DAA initiation and completion was confirmed by pharmacy records. Sustained virologic response (SVR) was defined as HCV RNA &amp;lt; 15 IU/ml measured &amp;gt;12 weeks after treatment completion. Results Of 282 patients scheduled for appointments, 84% (n=236) were white, 48% (134) were female, the median age of 39 years (interquartile range, 34-47) and 3% (11) were HIV-coinfected. Confirmatory labs were completed in 74% (210) of whom 23% (48) had spontaneously cleared HCV. Of 162 patients with detectable viremia, 92% (149) attended a provider visit, 90% (134) of those who attended a visit started treatment, 81% (108) of those who started treatment completed treatment, and 47% (51) of those who completed treatment had SVR labs drawn, of which 100% (51) achieved SVR (Figure 1). Overall, 45% (126) were prescribed buprenorphine. Conclusion We demonstrate an effective model for HCV care integration into rural SSPs with high uptake and completion of curative HCV treatment ( &amp;gt;80%). However, more than half did not complete post-treatment evaluation to confirm HCV response. Strategies for care after HCV treatment are needed. Disclosures Sherilyn Brinkley, MSN, CRNP, AbbVie: Honoraria|Gilead Sciences: Honoraria Mark S. Sulkowski, MD, AbbVie: Advisor/Consultant|Aligos Therapeutics: Advisor/Consultant|Gilead: Advisor/Consultant|GSK: Advisor/Consultant|GSK: Grant/Research Support|Janssen: Grant/Research Support|Precision Biosciences: Advisor/Consultant|Vir: Grant/Research Support|Virion: Advisor/Consultant Oluwaseun Falade-Nwulia, MBBS ,MPH, Abbvie Inc: Grant/Research Support|Gilead Sciences: Advisor/Consultant

Jumps of Jacobians via orthogonal canonical forms

Given a smooth, proper curve C C over a discretely valued field k k , we equip the k k -vector space H 0 ( C , ω C / k ) H^{0}(C,\omega _{C/k}) with a canonical discrete valuation v c a n v_\mathrm {can} which measures how canonical forms degenerate on regular integral models of C C . More precisely, v c a n v_\mathrm {can} maps a canonical form to the minimal value of its associated weight function, as introduced by Mustaţă–Nicaise [Algebr. Geom. 2 (2015), pp. 365–404]. Our main result states that v c a n v_\mathrm {can} computes Edixhoven’s jumps of the Jacobian of C C when evaluated in an orthogonal basis. As a byproduct, we deduce a short proof for the rationality of the jumps of Jacobians. We also show how v c a n v_\mathrm {can} and the jumps can be computed efficiently for the class of Δ v \Delta _v -regular curves introduced by Dokchitser [Duke Math. J. 170 (2021), pp. 2519–2574].

Energy–Carbon Coupling Modeling of Integrated Energy Systems in Low-Carbon Parks

In integrated energy system modeling, extant research predominantly addresses single-energy system optimization or carbon emission flow models, failing to adequately elucidate the mechanisms of combined energy and carbon flow modeling in complex energy systems. This deficiency hampers a thorough analysis of the coupling relationships between energy and carbon flows, thereby posing significant challenges for resource allocation and carbon mitigation within integrated energy systems. This paper presents an innovative energy–carbon coupling model, constructing a unified framework for energy and carbon flow modeling centered on the energy hub, thereby overcoming the limitations of traditional approaches that are unable to model both energy and carbon flows concurrently. The model comprehensively examines the coupling nodes and carbon density correlations among energy conversion devices within multi-energy systems, precisely quantifying carbon emission paths and distribution across devices. This provides a novel methodology for carbon emission management in integrated energy systems. Case studies on typical integrated energy systems demonstrate the proposed model’s efficacy in low-carbon economic dispatch. The energy–carbon coupling model developed in this study offers a high-adaptability solution for integrated energy systems in multi-energy, low-carbon parks, achieving an optimal balance between economic efficiency and environmental performance under dual objectives of energy demand and carbon emission minimization.

Predicting healthy lifestyle behavior among married people: integrating theory of planned behavior, health belief model, body image dissatisfaction and habit

PurposeThis study aimed to investigate the determinants of healthy lifestyle behavior in married people. It also sought to propose an integrative model that incorporates variables from two theoretical models, namely, the theory of planned behavior (TPB) and the health belief model (HBM), as well as two additional variables, namely habit and body image dissatisfaction.Design/methodology/approachQuestionnaires were used to gather data, and structural equation modeling (SEM) was applied for statistical analysis. Data were collected from 404 respondents using several assessment tools, including the Health Belief Model Scale, the Theory of Planned Behavior Scale, the Healthy Lifestyle Belief Scale, the Body Area Scale, and the Creature of Habit Scale.FindingsThe findings showed that perceived severity is the predictor of a healthy lifestyle attitude, followed by subjective norm, perceived behavior control, and healthy lifestyle attitude as the primary attendants of intention. Meanwhile, it was found that a healthy lifestyle attitude is a significant mediator in the perceived severity and healthy lifestyle intention relationship. The suggested model outperforms the original TPB and HBM models in predictive ability, proving its utility and efficacy in describing healthy lifestyle behavior.Originality/valueAdding constructs in the TPB and HBM models contributed considerably to improving the understanding of healthy lifestyle behavior among married people. Governments and policymakers might utilize the findings to design and execute a new program or a sustained healthy lifestyle for married people.

Social Determinants of Health and Adverse Childhood Experiences: How Are These Intertwined to Determine Childhood Overweight and Obesity in Argentina?

Argentina is one of the countries in the Latin American region with the highest incidence of childhood obesity. The incidence of childhood overweight and obesity among children under five years of age is approximately 15%, according to the 2022 MICS survey. Obesity is a nutritional disorder that is explained not only by biological factors but also by socioeconomic and family circumstances. The objective of this study is to investigate the factors and risks that determine childhood overweight and obesity in Argentine children under five years of age. To explain this, this article presents an integrative model of two theoretical frameworks, social determinants of health (SDH) and adverse childhood experiences (ACEs), called hereafter “IM-SDH-ACEs”, in order to understand how these determinants are related to influence childhood malnutrition. This study performs a Chi-square test to empirically corroborate that SDH and ACEs influence this type of weight disorder; thus, it is shown that socioeconomic status (SES) and age groups together with ACEs are factors associated with childhood overweight and obesity. Likewise, a multinomial logistic regression model is estimated, and the results show that the highest risks of suffering from overweight and obesity in childhood are the predictors of physical violence (PV) inflicted on the child as well as an interaction of the variables of deprivation in health care and a secondary education level of mothers or carers. It is concluded that a more inclusive social policy is required to reduce childhood obesity in Latin American societies.

Tell me more: integrating LLMs in a cultural heritage website for advanced information exploration support

Abstract Cultural Heritage websites’ capability to satisfy diverse information needs is limited by their high-quality but constrained knowledge bases. Thus, we investigate their extension with external large language models (LLMs), enriching the provision of cultural content by leveraging LLMs’ continuous collection and integration of information from heterogeneous data sources. This extension raises important challenges in synchronizing the LLM’s behavior with the user’s browsing activity on the website to offer a unified interaction environment. To address these challenges, we propose a loosely coupled integration model that provides users with curated content and an assisted question-answering function to answer information needs that the system’s knowledge base fails to cover. Our model is agnostic to the LLM and synchronizes its behavior with the user’s browsing activity through implicit prompt engineering. We tested a baseline website without LLM integration, one with free-text interaction with the LLM, and another that combines free-text interaction with the suggestion of context-dependent questions. In a user study involving 44 participants, we found that the LLM-powered website has higher usability and that context-dependent question suggestions further enhance user experience, especially for people with low curiosity levels (according to Curiosity and Exploration Inventory-II - CEI-II) who are guided in formulating effective questions. This shows the potential of LLMs to enrich engagement with existing Cultural Heritage websites.

An optimised and secured bio medical signal transmission using CNN

ABSTRACT Real-time transmission of biomedical signals remains challenging, particularly with IoT and cloud-based solutions, due to cybersecurity risks and energy constraints. To ensure secure and power-efficient transmission, we propose a selective transmission-based reconstruction model. Instead of sending the entire signal, only the most relevant features are transmitted to the intended receiver. At the destination, an encrypted version of the signal is reconstructed using a decryption algorithm, preventing data loss. The primary challenge lies in sensor network monitoring, especially regarding energy consumption and data acquisition. To address this, we integrate compressive sensing methods to enhance data reconstruction and prolong network lifetime. This paper introduces the CNN-based Optimized BCAE, an integrated model combining an autoencoder with feature selection, allowing effective extraction of key features for biomedical applications. By distinguishing relevant from irrelevant task units, this approach improves data throughput and optimizes power management. Our model offers a reliable solution for real-time biomedical signal processing, ensuring both efficiency and security.

Hybrid learning strategies: integrating supervised and reinforcement techniques for railway wheel wear management with limited measurement data

Train wheel wear significantly impacts wheel-rail interaction forces and is an unavoidable issue in the railway industry. This study focuses on regular wear, specifically changes in wheel profiles such as tread wear, flange height, and flange thickness. Effective wheel wear management is crucial for maintaining the reliability, safety, and efficiency of rail systems. However, regular measurement of wheel profiles is often limited by constraints such as dense traffic, budget, time, and remote assets, which reduces the effectiveness of traditional maintenance approaches. This study proposes a hybrid learning strategy combining supervised and reinforcement learning techniques to optimize train wheel wear management under these constraints and achieve predictive maintenance. The supervised learning model, developed from validated simulations, predicts wear progression, while reinforcement learning improves maintenance decision-making using basic operational data without regular measurements. Various machine-learning techniques are explored and fine-tuned to identify the best models for preventing faulty wheels without the need for frequent inspections. By integrating these two learning approaches, the framework enhances the accuracy of wear predictions and optimizes maintenance schedules, reducing the risk of over-maintenance or unexpected failures. This integrated model addresses challenges such as system complexity, limited data, and cost-effectiveness in the industry. In terms of supervised learning, the R2 for tread wear prediction improves from 0.94 to 0.95 compared to previous studies, and the model, when integrated with reinforcement learning, significantly reduces defects based on wear and irregular wheel dimensions. This research is the first to integrate supervised and reinforcement learning specifically for train wheel wear management under limited measurement data constraints, offering a breakthrough compared to traditional methods that rely on regular inspections. The study provides significant benefits for the railway industry, including reduced maintenance costs, improved maintenance efficiency, lower defect rates, reduced possession and inspection time, and enhanced passenger comfort and safety.