Standard Machine Research Articles

Ranking approaches for similarity-based web element location

Context:GUI-based tests for web applications are frequently broken by fragility, i.e. regression tests fail due to changing properties of the web elements. The most influential factor for fragility are the locators used in the scripts, i.e. the means of identifying the elements of the GUI. Objective:We extend a state-of-the-art Multi-Locator solution that considers 14 locators from the DOM model of a web application, and identifies overlapping nodes in the DOM tree (VON-Similo). We augment the approach with standard Machine Learning and Learning to Rank (LTR) approaches to aid the location of web elements. Method:We document an experiment with a ground truth of 1163 web element pairs, taken from different releases of 40 web applications, to compare the robustness of the algorithms to locator weight change, and the performance of LTR approaches in terms of MeanRank and PctAtN. Results:Using LTR algorithms, we obtain a maximum probability of finding the correct target at the first position of 88.4% (lowest 82.57%), and among the first three positions of 94.79% (lowest 91.86%). The best mean rank of the correct candidate is 1.57. Conclusion:The similarity-based approach proved to be highly dependable in the context of web application testing, where a low percentage of matching errors can still be accepted.

Evolving Neural Network Designs with Genetic Algorithms: Applications in Image Classification, NLP, and Reinforcement Learning

A method of evolving deep learning architectures using genetic algorithms is presented. The method is a first step towards a low-cost evolutionary search for task-specific neural networks. We evolve task-specific model architectures optimized for fast execution and low error on several standard machine learning tasks: image classification, character-level language modeling, and solving the cart pole problem. We also introduce a simple variation of the method that is capable of evolving neural networks with recurrent connections of varying depth and length and show performance on a word-level language modeling task. The method is implemented in an open-source library. We hope that the ability to run an evolutionary search at this scale will make it possible for a wide audience to develop deep learning architectures that are specialized for a variety of tasks and to develop many interesting novel architectural features. A new method that uses evolutionary search to directly modify existing neural network architectures to perform a specific task is presented. We demonstrate that task-specific specialization of deep learning models can be useful in practice. We modify convolutional neural networks, residual networks, and an LSTM variant to perform various tasks, and show that specialized networks often perform better than models trained from scratch that have many more parameters and much larger training time. For example, on the object recognition task, a specialized model is built by training a base network to predict object position and then applying a series of genetic search operations to squeeze the network and fit new final layer weights to the output. The specialized model is 8 times faster and has 13% lower error, despite being 17 times smaller than a fully trained larger and slower network.

Is Earthquake Psychological Trauma a Cause of Post-Earthquake Dizziness? (Using Standard Vector Machine)

The aim of this study is to reveal the relationship between complaints such as dizziness of university students who experienced an earthquake in Kahramanmaraş and surrounding provinces on February 6, 2023 and the psychological trauma caused by the earthquake. 476 university students who survived the earthquake that occurred in Kahramanmaraş on February 6, 2023 in Turkey were included in this study. In the first part of the survey, students were asked about dizziness that did not occur before the earthquake but occurred after the earthquake, and in the second part of the survey, a total of 41 questions were asked in the six sub-dimensions of the attitude scale to determine the psychological states of individuals exposed to the earthquake, and each of the six variables, namely detachment from life, social health, psychological change, trauma anxiety, maturation, and avoidance, was compared with dizziness and the relationship between dizziness and the other six variables was tried to be found. In this study, the ‘’Attitude Scale for Determining the Psychological States of Individuals Exposed to Earthquakes’’ was used to reveal the psychological states of the students; and the ‘’Standard Vector Machine’’ (VRM) was used for statistical studies. The study concluded that dizziness that did not occur before the earthquake but occurred after the earthquake was closely related to the psychological trauma experienced due to the earthquake.

Unified Machine Learning Techniques for High-Dimensional Survival Data Analysis

Survival analysis is a statistical method used to analyze time-to-event data. The existence of states where event results become invisible after a given time is one of the primary difficulties with this regulation. This is considered censorship. This kind of data may be found in a variety of applications, such as mechanical system failure rates, patient mortality rates during clinical trials, and the length of unemployment in a community. Assessing the so-called survival and hazard functions is one of the primary goals of survival analysis. Many machine-learning algorithms have also been developed to tackle censored data and other difficult problems with real-world data. Machine-learning techniques enhance survival analysis by incorporating flexible modeling approaches, handling high-dimensional data, capturing nonlinear relationships, and accounting for censorship. They provide powerful tools to extract meaningful insights and make accurate predictions in time-to-event analysis across various healthcare, finance, and engineering domains. The standard statistical approaches and machine learning techniques advanced for survival analysis are structured in this unified framework, as is the implementation of some machine learning techniques applying to the GBSG2 survival analysis dataset.

Barriers and facilitators to point-of-care ultrasound use in an academic emergency department by perceived usability

BackgroundWhile use of point-of-care ultrasound (POCUS) has become widespread in emergency medicine, its adoption and usage among emergency clinicians is variable. In this study, we explored the barriers and facilitators to POCUS use among emergency medicine clinicians in a tertiary care emergency department in the United States by clinical role and perceived usability of POCUS. MethodsWe initially administered a quantitative survey via REDCap and used a validated technology usability scale to categorize clinicians into tertiles of low, moderate, and high, based on perceived utility of POCUS. Subsequently, a purposeful sample in each category participated in semi-structured interviews until thematic saturation was reached (9–10 per category, 29 total). The qualitative interviews were analyzed using Braun and Clark's content analysis approach. A coding scheme was developed through initial review of 20 % of the transcripts, followed by operational definitions of codes, concept and category development, and selection of representative quotes. Analysis was supported by Dedoose software. ResultsWe identified 41 % (12/29) as high POCUS users, 31 % (9/29) as moderate, and 28 % (8/29) as low, based on perceived usability. Interviewees consisted of 24 % (7/29) attendings, 28 % (8/29) residents and 48 % (14/29) advanced practice providers (APPs). Attendings made up 50 % in the high POCUS group and 75 % of the low POCUS group were APPs. In the low group, the lack of prior and present POCUS training was a major barrier to POCUS use. The lack of time during clinical shifts was a major barrier to POCUS utility in the high and low groups. Availability of POCUS machines in clinical spaces was identified as a facilitator to POCUS use in the medium group. ConclusionsParticipants reported that POCUS facilitates patient disposition and clinical supervision enhances its use. Early POCUS education in professional school and continued POCUS training in clinical practice could facilitate POCUS use clinically. Structured POCUS courses and continued medical education programs may provide protected time to learn and practice POCUS. Moreover, accessible and standardized machines in the clinical environment could improve POCUS usage.

Impact of Target Surface Building Direction on the Heat Transfer Characteristics of Additive Manufactured Impingement Systems

Additive manufacturing (AM) is widely recognized as a prominent tool to maximize the potential of internal cooling systems for gas turbine applications. Several past studies have been undertaken in order to assess the effect of additive manufactured components peculiar characteristics, mainly in the form of surface roughness, on heat transfer and pressure losses. On the other hand, impingement constitutes one of the most adopted solutions for turbine vane internal cooling; also, its heat transfer performance has been shown to be potentially improved through the use of roughened target surfaces in several studies. In this work, the effect of AM-generated roughness on the performance of impingement systems has been experimentally investigated. A lumped approach was used to test additive manufactured coupons reproducing an impingement array in 1:1 scale and retrieve an average heat transfer assessment. The Laser Powder Bed Fusion (L-PBF) technique was used for the manufacturing process. As one of the main parameters affecting AM-generated roughness, the building direction of the target surface was varied in order to highlight its impact on the overall performance comparing four different building directions with a smooth reference target plate made by standard CNC machining.

Artificial Intelligence Vision Methods for Robotic Harvesting of Edible Flowers.

Edible flowers, with their increasing demand in the market, face a challenge in labor-intensive hand-picking practices, hindering their attractiveness for growers. This study explores the application of artificial intelligence vision for robotic harvesting, focusing on the fundamental elements: detection, pose estimation, and plucking point estimation. The objective was to assess the adaptability of this technology across various species and varieties of edible flowers. The developed computer vision framework utilizes YOLOv5 for 2D flower detection and leverages the zero-shot capabilities of the Segmentation Anything Model for extracting points of interest from a 3D point cloud, facilitating 3D space flower localization. Additionally, we provide a pose estimation method, a key factor in plucking point identification. The plucking point is determined through a linear regression correlating flower diameter with the height of the plucking point. The results showed effective 2D detection. Further, the zero-shot and standard machine learning techniques employed achieved promising 3D localization, pose estimation, and plucking point estimation.

Computer Vision Foundation Models in Endoscopy: Proof of Concept in Oropharyngeal Cancer.

To evaluate the performance of vision transformer-derived image embeddings for distinguishing between normal and neoplastic tissues in the oropharynx and to investigate the potential of computer vision (CV) foundation models in medical imaging. Computational study using endoscopic frames with a focus on the application of a self-supervised vision transformer model (DINOv2) for tissue classification. High-definition endoscopic images were used to extract image patches that were then normalized and processed using the DINOv2 model to obtain embeddings. These embeddings served as input for a standard support vector machine (SVM) to classify the tissues as neoplastic or normal. The model's discriminative performance was validated using an 80-20 train-validation split. From 38 endoscopic NBI videos, 327 image patches were analyzed. The classification results in the validation cohort demonstrated high accuracy (92%) and precision (89%), with a perfect recall (100%) and an F1-score of 94%. The receiver operating characteristic (ROC) curve yielded an area under the curve (AUC) of 0.96. The use of large vision model-derived embeddings effectively differentiated between neoplastic and normal oropharyngeal tissues. This study supports the feasibility of employing CV foundation models like DINOv2 in the endoscopic evaluation of mucosal lesions, potentially augmenting diagnostic precision in Otorhinolaryngology. 4 Laryngoscope, 134:4535-4541, 2024.

Reduction of the residual warpage of fused deposition modeling by negative thermal expansion filler

We investigated the suppression of residual warpage using a negative thermal expansion (NTE) material filler, Zn1.6Mg0.4P2O7. To focus on the effect of thermal expansion on residual warpage, we used acrylonitrile-butadiene-styrene, which is an amorphous polymer with a small volume change around the liquid–solid phase transition, as the matrix. Composite pellets were prepared using a kneader, and filaments were produced for fused deposition modeling (FDM) using an extruder. We fabricated a bar-like test piece using a standard FDM machine and measured the warpage deformation using 3D scanning. The experimental results were supported by the finite element method. We also compared similarly sized SiO2 powders to discuss the advantages of NTE over conventional fillers. The warpage of a 100 mm x 5 mm x 2 mm bar-like test piece was reduced by approximately 75% by introducing Zn1.6Mg0.4P2O7 at 40 vol%. The effect of Zn1.6Mg0.4P2O7 at 30 vol% is similar to that of a conventional SiO2 filler at 40 vol%. Reducing the residual deformation through fillers would reduce energy consumption and simplify the device by avoiding build stage and chamber heating.

Central Line-Associated Bloodstream Infection Reduction in Hemodialysis Patients Across 9 Hospitals and 3 States.

This quality improvement project was initiated to reduce hospital-acquired catheter-associated bloodstream infections (CLABSI) in hospitalized patients receiving dialysis. A team dedicated to reducing hospital-acquired infections led the implementation of evidence-based interventions across all the included hospitals. This innovative approach demonstrated substantial enhancements in outcomes for patients on hemodialysis. To enhance patient safety in patients receiving hemodialysis, new strategies were implemented, including (1) transitioning from a vendor model to an internal model, enabling dialysis program standardization, (2) empowering intensive care nurses with increased autonomy and ownership, (3) transitioning to a standardized dialysis machine, and (4) introducing chlorhexidine gluconate (CHG) impregnated caps and CHG pads. To reduce CLABSI in hemodialysis lines, a multidisciplinary team was formed comprising physicians, nurses, a dialysis technician, pharmacists, the dialysis director, the chief medical officer, the chief nursing officer, the assistant chief nursing officer, the infection preventionist, and the quality director. The team implemented a standardized approach to caring for hemodialysis lines, provided just-in-time education to staff, and standardized policies simultaneously at 8 hospitals. Initially, 1 facility served as the pilot facility for facility-owned dialysis services, totaling 9 facilities providing in-house standardized dialysis services. Data was reported back for monthly evaluation. Overall, there was an 88% reduction in CLABSI occurrences in hemodialysis lines from pre-intervention (n = 8) to post-intervention (n = 1), X2 (1, N = 4112) = 4.181, P = .0408. Collaboration on these initiatives improved communication and enhanced quality care and patient safety across the entire spectrum of care. Implementing innovative tracking of standardized approaches to patient care and infection prevention and evidence-based interventions resulted in decreased CLABSI rates, improving outcomes in vulnerable patients. An unintended benefit of this project was the increase in multidisciplinary collaboration.

Analysis of the mean electrical axis of the heart from standard electrographic leads in normal adult males

Background: Electrocardiogram has been widely applied as a diagnostic tool for cardiovascular disease. The concept of the cardiac vector is crucial for the clinical application in ECG interpretation. Methods: A prospective cross-sectional study was conducted among 116 male subjects, aged 18-25 years, comprising students and staff of Gandaki Medical College Teaching Hospital and Research Center, Pokhara, Nepal, over the period from June to September 2023. Electrocardiograms (ECGs) were recorded using a standard ECG machine with conventional limb leads. The mean electrical axis of the heart for each individual was determined by plotting the net voltage of the QRS complex in Lead-I and Lead-III. The study aimed to analyze the correlation between the cardiac vector and physical measurements, such as height, weight, Body Surface Area (BSA), and Body Mass Index (BMI). Data analysis was performed using SPSS version 27.0. Results: The normal mean electrical axis of the healthy male subjects was observed as 57.88±24.55°. There was significant positive correlation of cardiac vector with height (p< 0.05), whereas negative correlation was observed with weight and BMI (p< 0.01). However, there was no significant correlation with BSA. In our study we observed the maximum left axis cardiac vector as -10° and right axis as 90° among 116 male subjects. Conclusion: We established a cardiac vector with values differing from those reported in other studies, using standard bipolar limb leads in normal, healthy male subjects. It is also examined that greater BMI has shifted the electrical activity of the ventricles to the left.

A mcgyvering way to left bundle branch area pacing : using the treadmill test machine ECG system and modified V1 / V6 leads connected to the pacing system analyser for left bundle branch area pacing

Abstract Background Left bundle branch area pacing (LBBaP) is a novel pacing modality which provides physiological synchronisation of both the ventricles. However, its use and adoption requires the presence of a standardised electrophysiological (EP) system with 12 lead ECG to show left bundle branch area capture. Purpose The aim of the study was to assess the feasibility and outcomes of using the conventional treadmill test (TMT) machine 12 lead ECG system and modified V1 and V6 leads connected to the pacing system analyser (PSA) to demonstrate conduction system capture during left bundle branch area pacing (LBBaP). Methods LBBaP was attempted by a single operator in a mixed cohort of patients at peripheral hospitals lacking the standardised EP system. The presence of LBB area capture was tested with the aid of the 12 lead ECG system connected to the standard TMT machine (1). Conduction system capture was also assessed using modified V1 and V6 leads connected to the PSA for measurement of V6 left ventricular activation time (LVAT)(2) . The baseline clinical, procedural, pacing, electrocardiographic parameters, echocardiographic response (> 10% improvement in ejection fraction {EF}) and clinical response to LBBaP were assessed. Results LBBaP was attempted in 19 patients and was successful in 18 patients (94.7%). Patient characteristics include age 66.7 +/- 11.47 years, female 47.37%, male 52.63%, ischaemic cardiomyopathy 10.52% and non-ischaemic cardiomyopathy 5.26% .The mean baseline EF and QRS duration was 45%+/-14% and 130.18 +/-10.45 milliseconds(ms) respectively. LBBaP threshold was 0.59 +/-0.168 volts and R wave amplitude was 12+/-3.5 millimetres , which remained stable during the mean follow up of 15.9 +/-13.67 months. The mean fluoroscopic duration was 11.57 +/-2.58 minutes and procedural time was 108.78+/-29.20 minutes. LBBaP resulted in impressive narrowing of the QRS duration to 102.39 +/-10.55 ms (p<0.001) with LVAT of 68.94 +/-9.94 ms. LVEF improved in the patients with cardiomyopathy from 33.4 +/- 5.77 % to 48.2+/-12.37 % (p=0.028) .Clinical and echocardiographic response was noted in 87.5% of patients with cardiomyopathy. Conclusion The 12 lead ECG system with the standardised TMT machine in conjunction with the modified V1 and V6 leads connected to the PSA is a feasible alternative to the EP system to perform successful left bundle branch area pacing.ECG leads connected during LBBapClinical profile of patients

Handheld ultrasound versus standard machines for placement of peripheral IV catheters: A randomized, non-inferiority study

IntroductionUltrasound guided IV catheter (USGIV) access occurs frequently in Emergency Departments (EDs). This task is often performed using large, expensive, cart-based ultrasound systems (CBUS) which are frequently needed for other ED ultrasound functions and can be cumbersome to use and store. Handheld ultrasounds (HHUs) may be able to meet this need, but it is unknown if they function interchangeably with CBUS for USGIV placement. We performed a prospective, randomized, noninferiority study to compare the success rate of HHUs to CBUSs for placing USGIVs. MethodsED patients 18 and older needing an USGIV were approached for enrollment and randomized to receive an USGIV placed by CBUS or HHU. USGIVs were placed by any ED physician or nurse trained in placement. A placement was considered attempted upon needle entry into the skin. An USGIV was successful if it was immediately flushable with saline. Data was collected on the success of IV placement, number of attempts, IV and provider characteristics, patient demographics, and length of time the USGIV lasted. Demographics and operator and IV characteristics were analyzed using Pearson's Chi square, Fischer's Exact test, or Wilcoxon rank sum tests. Non-inferiority was assessed using the Farrington-Manning test. Results were approached per protocol and analyzed in R. Results312 patients were enrolled. Patient and IV characteristics were similar between groups. There was no difference in the number of successful USGIVs placed in either group (p≥0.9) with 146 in the CBUS group and 145 in the HHU group. There was no difference in the first attempt success rate between groups (p = 0.8) and HHU was noninferior to CBUS for successful USGIV placement (p = 0.0001). The rate of premature USGIV failure was similar between HHU and CBUS (4.0 % and 6.7 %). ConclusionHHU was noninferior to CBUS for successful USGIV placement. There was no difference in the rate of first attempt success at placement or USGIV survival to a patient's ED disposition between groups. No significant additional training was required for ED providers of all levels to use the HHUs.

Development of an Artificial Intelligence System for Distinguishing Malignant from Benign Soft Tissue Tumors Using Contrast-Enhanced MR Images.

The integration of artificial intelligence (AI) into orthopedics has enhanced the diagnosis of various conditions; however, its use in diagnosing soft-tissue tumors remains limited owing to its complexity. This study aimed to develop and assess an AI-driven diagnostic support system for magnetic resonance imaging (MRI)-based soft-tissue tumor diagnosis, potentially improving accuracy and aiding radiologists and orthopedic surgeons. Experienced orthopedic oncologists and radiologists annotated 720 images from 77 cases (41 benign and 36 malignant soft-tissue tumors). Eleven tumor subtypes were identified and classified into benign and malignant groups based on histological diagnosis. Utilizing the standard machine learning classifier pipeline, we examined and down-selected imaging protocols and their predominant radiomic features within the tumor's three-dimensional region to differentiate between benign and malignant tumors. Among the scan protocols, contrast-enhanced T1 weighted fat-suppressed images showed the most accurate classification based on radiomics features. We focused on the two-dimensional features from the largest tumor boundary surface and its neighboring slices, leveraging texture-based radiomic and deep convolutional neural network features from a pretrained VGG19 model. The test data comprised 44 contrast-enhanced images (22 benign and 22 malignant soft-tissue tumors) containing six malignant and five benign subtypes distinct from the training data. We compared expert and non-expert human performances against AI by assessing malignancy detection and the time required for classification. The AI model showed comparable accuracy (AUC 0.91) to that of radiologists (AUC 0.83) and orthopedic surgeons (AUC 0.73). Notably, the AI model processed data approximately 400 times faster than its human counterparts, showcasing its capacity to significantly boost diagnostic efficiency. We developed an AI-driven diagnostic support system for MRI-based soft-tissue tumor diagnosis. While additional refinement is necessary for clinical applications, our system has exhibited promising potential in differentiating between benign and malignant soft-tissue tumors based on MRI.

Reuse of polyvinyl chloride residues—in powder and fiber forms—as potential components for green mortar pavers

Plastic wastes are considered a significant environmental pollution load because of their mostly non-biodegradable nature. Consequently, the need to lessen the environmental impact calls for employing eco-friendly strategies such as reusing materials in construction. The objective of this study is to analyze the mechanical behavior and environmental impacts of mortar elements, with a specific focus on pavers, by incorporating waste polyvinyl chloride (WPVC) as aggregate and partially replacing cement. Two distinct forms of this waste were considered in the investigation: (i) as it is generated in industrial processes, mainly in fiber form, WPVCf, and in the form of powder obtained by sieving the former, WPVCp. The ratio of conventional mortar materials replaced by WPVC and the final water-to-binder ratio changed based on waste type and mix workability. The experimental work involved preparing 84 pavers, each measuring 100 mm × 150 mm × 60 mm. For each type of paver, 12 samples were prepared, ensuring a minimum of 4 sets of test data for each curing period (7, 14, and 28 days). Flexural strength tests, conducted in accordance with the Colombian standard NTC 2017, utilized a standard flexural test machine with a 150 kN capacity. A specially designed steel tool applied a point load at the center of each paving block, maintaining a minimum distance of 10 millimeters from the edges. The findings demonstrated that pavers containing fiber plastic waste exhibited the requisite strength for paved walkways, even in the short term, when the weight ratio of WPVCf to sand (WPVCf/S) was 24 %. These results derive an eco-friendly solution for reusing PVC wastes currently deposited directly in landfills. Moreover, to analyze the environmental impacts of this solution, a Life Cycle Assessment (LCA) analysis was carried out in this research. The LCA results pointed out that dosing with a WPVCf/S ratio of 24 % emerges as the most ecologically responsible, manifesting substantial reductions across impact categories, such as a significant 23 % decrease in climate change, a 23 % reduction in Photochemical ozone formation, and a 22 % reduction in renewable energy consumption. This underscores the potential of WPVCf incorporation as a consequential step toward eco-friendly construction materials. Furthermore, this solution not only translates to diminished paving block costs by requiring less fine aggregate and cement but also contributes to reduced transportation expenses owing to the decreased weight of the pavers.

An ensemble deep learning model for fast classification of Twitter spam

Twitter spam needs to be detected and arrested quickly. The paper examines methods for classification of spam in terms of determination of important features, comparative performance of classification models, and improvement in time performance for classification. It presents a conceptualization of several novel rich, deep, and naïve features. The extraction processes for rich and deep features increase the time complexity of spam classification. To address this, the proposed model selectively segregates and combines features to enable near real-time processing. This supersedes the time performance of standard machine learning and deep learning models, with no compromise on the quality of classification.

Impact sound insulation performance measurements on CLT floors – Effect of excitation source type: Tapping machine and rubber ball

Laboratory and field measurements of the impact sound acoustic performance of a CLT wood-based floors have been performed. French national regulation requires the use of the standardized tapping machine for impact sound performance measurements. Since the early 2000s, rubber ball impact source has been standardized in some Asian countries to measure and evaluate the low frequency impact sound performance of floors; this excitation source is now part of international standards concerning building acoustic measurement in laboratories and in-situ. Indeed, the impact sound level associated to this soft impact source is supposed to provide a better correlation with annoyance from jumping and/or running children in buildings. The paper reviews the measurements performed using the standard tapping machine and the rubber ball as impact source, on CLT based floors in a laboratory, as well as those on the CLT based experimental building. Impact sound improvements are presented and discussed. The necessity of measuring with both impact sources, rather than only with the tapping machine is examined.

Mathematical Modelling and Deep Learning Techniques for Predicting Cardiovascular Disease

Putting together mathematical models and deep learning has become a strong way to identify cardiovascular disease (CVD) in recent years. This paper looks at how dynamical systems, control theory, and machine learning methods can work together to make CVD forecast more accurate and reliable. To improve the ability of diagnostic tools to predict the future by using mathematical models like differential equations to describe how the heart works and deep learning architectures like Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) for feature extraction. This method considers both the time patterns of how the heart works and the non-linear complexity of risk factors like high blood pressure, high cholesterol, and family tendencies. The most important part of our method is creating a multi-stage prediction model. This model starts with differential equations that show how heart rate and blood pressure change over time. Next, it uses deep learning models that have been trained on large CVD datasets. Normalization and feature engineering methods are used to prepare the information. Key measures like age, BMI, and cholesterol levels are used as inputs. Metrics like precision, sensitivity, specificity, and F1-score are used to judge how well the model works. Early findings indicate that combining mathematical models with deep learning makes predictions a lot more accurate than using only standard machine learning models. In particular, adding time dynamics to the model lets it find hidden trends in physiological data that simple models often miss. Deep learning also lets the system learn on its own how different risk factors are related to each other, making it a strong tool for finding CVD early and predicting how bad it will be. The research shows that a mix of approaches could make personalized medicine better by giving doctors accurate and easy to understand prediction tools. In the future, researchers will focus on making these models work better so it can be used in real time in hospital settings. This could help lower the number of people who get cardiovascular disease around the world.

Computational prediction of impact sound insulation of a timber floor excited by an ISO standard tapping machine

During the last decades, researchers have actively developed procedures to predict impact sound insulation of timber slabs and floors using the finite element method (FEM). Since the simulations have mainly been performed in the low frequencies, a full understanding of their suitability for research and development (R&D) purposes has not yet been created. The object of the study reported in this paper was to determine the applicability of a modern FEM simulation procedure for predicting the normalised impact sound pressure level Ln of a full timber floor, when information provided by material manufacturers is used as input data. Force excitation caused by the ISO standard tapping machine was determined utilizing explicit time integration and FEM. The modelling of sound radiation itself was performed in a frequency domain FEM analysis. The models of a timber floor in its construction stages were constructed before the laboratory measurements of impact sound insulation so that the set-up corresponded to a simulated R&D task. Based on the prediction results, the applied procedure was considered suitable for R&D purposes and needs for further studies were identified.

Developing guidelines for a reproducible method for low-frequency impact measurements in buildings

The impact noise performance in buildings is evaluated using the measured radiated sound pressure level when impacted by a standard tapping machine. Recent work has shown that there can be 4 - 5 dB variation in the sound pressure levels at different measurement locations in the room. This can lead to a high reproducibility error for field tests. In this work, the authors have developed a novel reciprocity-based measurement method that shows approximately 1 - 1 .25 dB standard deviation in low-frequencies. Additionally, this method provides output/input ratio-based measurements, which can be useful in design predictions. The method was developed using a simulation model and verified using testing in three structures.