Low Performance Research Articles

IoT-HGDS: Internet of Things integrated machine learning based hazardous gases detection system for smart kitchen

This paper proposes an Internet of Things (IoT) and Machine Learning (ML) integrated Hazardous Gas Detection System (IoT-HGDS) for smart kitchens. The design incorporates six tin-oxide-based metal–oxide–semiconductor (MOS) gas sensor arrays and one DHT22 (temperature & humidity sensor). This IoT-HGDS can detect different hazardous gases, Volatile Organic Compounds (VOCs), and odors responses released from the kitchen’s materials and transmit them to a Remote Data Processing Centre (RDPC) through Amazon-Web Services (AWS) in real time. In this experiment, we collected 150×9=1350 samples from 9 kitchen materials like ghee, milk, liquid petroleum gas (LPG), bread, mustard oil, compressed natural gas (CNG), pigeon peas, refined oil, and kerosene. The Standardized Independent Component Analysis (SICA) pre-processing technique has been used to clean data, standardise the features, and remove outliers. ML approaches like Logistic Regression (LR), Adaptive Boosting (AdaBoost) and Regularized Discriminant Analysis (RDA) have been applied for accurate identification of gases/VOCs class and provide immediate alerts to improve kitchen safety. The SICA-RDA classifier outperformed (highest accuracy at 97.78 %) as compared to LR and AdaBoost in terms of performance and balanced precision, recall, and F1-Score. LR has the lowest performance in all metrics. LPG has the lowest Mean Squared Error (MSE) of 6.62×10−7, while CNG has the highest MSE of 3.60×10−4. This system can intelligently preserve gases, ensure safety precautions, and prevent accidents in the kitchens.

The relationship between lead levels and reading acquisition in Spanish speakers, evidence from Uruguayan schoolers.

To investigate the cross-sectional association between blood lead levels (BLL), and pre-reading and reading abilities in first-grade children from Montevideo, Uruguay. Of 357 school children (age 67-105 months) enrolled into the study, 287(43 % girls) had a BLL measure and an assessment of pre-reading and reading abilities based on five tests (Verbal comprehension, Sound blending, Letter word identification, Sentence reading fluency, and Passage comprehension) from the Batería III Woodcock-Muñoz. Separate generalized linear models (GLM) were conducted on the relationship between BLL and each test score separately, adjusting for sex, maternal education, household assets, Home Observation for Measurement of the Environment Inventory score, season, test administrator, blood lead testing method, and school clusters. The mean BLL was 4.0 ± 2.2 µg/dL, with no differences between the sexes. BLL was associated with a poorer vocabulary knowledge (β [95 % CI]): -0.20 [-0.39, 0.01]. For all the tests, children with BLLs ≥5 µg/dL tended to exhibit poorer performance than children with lower BLLs, but these associations were not statistically significant. When stratified by sex, some evidence of differential associations between BLLs and reading abilities emerged: BLLs were associated with higher phonological awareness in girls (0.32 [0.15, 0.48]) but not boys, and with lower reading comprehension in boys (-0.54 [-1.20, 0.13]) but not girls. Also, lead exposure (BLL ≥ 5 µg/dL) was more strongly and negatively associated with phonological awareness (-1.22 [-1.57, -0.86]) in boys than girls. In this study of first-grade children learning to read in Spanish, we found an inverse association between lead exposure and vocabulary scores, as well as tendency toward lower performance on pre-reading and reading measures among children with BLLs ≥5 µg/dL. Pre-reading and reading abilities are relevant to literacy acquisition; further research is required to confirm these links in larger studies, and to investigate differences between boys and girls, and according to key sociodemographic characteristics.

Facile fabrication of sulfonated bacterial cellulose-silane composite aerogels via in suit polymerization for enhanced oil/water separation performance.

Cellulose-based aerogels have emerged as highly promising materials for oil-water separation because of their highly porous nature, low bulk density, cost-effectiveness, and functional performance. In this study, a novel, robust, and hydrophobic bacterial cellulose aerogel (HBCA) was reported for highly efficient oil/water separation, which was formed by incorporating methyltrimethoxysilane (MTMS) into sulfonated nano-fibrillated bacterial cellulose (SNBC) matrix through freeze-drying method. The structural integrity of the SNBC/MTMS aerogel was ensured by its three-dimensional linked network structure. The resultant aerogel demonstrated superior hydrophobicity with a contact angle of 152.4°. As an oil-absorbing material, it selectively adsorbed different types of oils and organic solvents, with a saturation adsorption capacity ranging from 42.14 to 85.37 g·g-1. Additionally, this composite aerogel demonstrated outstanding separation efficiency (98.48 %) in continuous oil-water mixture processes, suggesting promising potential applications in the treatment of industrial oil pollution and oil spill accidents. This study indicates that the proposed HBCA with sulfonated nano-fibrillated bacterial cellulose is a potential and effective material for addressing environmental challenges related to oil contamination.

Designing and additive manufacturing of biomimetic interpenetrating phase zirconia-resin composite dental restorations with TPMS structure

Zirconia and resin are the most commonly utilized materials in dental restorations. However, zirconia presents significant wear on opposing teeth, whereas resin materials have low wear resistance and mechanical performances. A zirconia-resin interpenetrating phase composite (IPC) dental restoration was designed and fabricated using 3D printing and vacuum infiltration processes, incorporating zirconia scaffolds with triply periodic minimal surfaces (TPMS) structures. The mechanical and tribological performances of the IPCs were investigated through compressive and tribological experiments and finite element analysis, elucidating the influence of zirconia volumetric fraction. Results showed that IPCs exhibit excellent mechanical and tribological compatibilities, which can reduce the damage and wear of the antagonistic teeth. This designing and manufacturing strategy enables the IPC restorations with promising applications in dentistry.

Effects of mixed metal exposures on MRI diffusion features in the medial temporal lobe

BackgroundEnvironmental exposure to metal mixtures is common and may be associated with increased risk for neurodegenerative disorders including Alzheimer’s disease. This study examined associations of mixed metal exposures with medial temporal lobe (MTL) MRI structural metrics and neuropsychological performance. MethodsMetal exposure history, whole blood metal, MRI R1 (1/T1) and R2* (1/T2*) metrics (estimates of brain Mn and Fe, respectively), and neuropsychological tests were obtained from subjects with/without a history of mixed metal exposure from welding fumes (42 exposed subjects; 31 controls). MTL structures (hippocampus, entorhinal and parahippocampal cortices) were assessed by morphologic (volume or cortical thickness) and diffusion tensor imaging [mean (MD), axial (AxD), radial diffusivity (RD), and fractional anisotropy (FA)] metrics. In exposed subjects, effects of mixed metal exposure on MTL structural and neuropsychological metrics were examined. ResultsCompared to controls, exposed subjects displayed higher MD, AxD, and RD throughout all MTL ROIs (p’s<0.001) with no morphological differences. They also had poorer performance in processing/psychomotor speed, executive, and visuospatial domains (p’s<0.046). Long-term mixed metal exposure history indirectly predicted lower processing speed performance via lower parahippocampal FA (p’s<0.023). Higher entorhinal R1 and whole blood Mn and Cu levels predicted higher entorhinal diffusivity (p’s<0.043) and lower Delayed Story Recall performance (p=0.007). DiscussionMixed metal exposure predicted certain MTL structural and neuropsychological features that are similar to those detected in Alzheimer’s disease at-risk populations. These data warrant follow-up as they may illuminate a potential path for environmental exposure to brain changes associated with Alzheimer’s disease-related health outcomes.

Comprehensive Performance Evaluation of Machine Learning Algorithms for detecting DDoS attacks in SDN

&lt;div align="center"&gt;Software-Defined Networking (SDN) revolutionizes networking by separating control logic and data forwarding, enhancing security against threats like Distributed Denial of Service (DDoS) attacks. These attacks flood control plane bandwidth, causing SDN network failures. Recent studies emphasize the efficacy of machine learning and statistical approaches in identifying and mitigating these security risks. However, there has been a lack of focus on employing ensembling techniques, amalgamating diverse machine learning models, selecting pertinent features, and utilizing oversampling techniques to balance categorical data. Our study evaluates 20 machine-learning models, emphasizing feature engineering and addressing class imbalance using Synthetic Minority Oversampling TEchnique (SMOTE). The results indicate that Ensemble methods such as LGBM Classifier, Random Forest Classifier, XGB Classifier, Decision Tree Classifier attained near-perfect scores (almost 100%) across all metrics, suggesting potential overfitting. Conversely, models like AdaBoost Classifier, K-Neighbors Classifier, and SVC exhibited slightly lower (99%) but realistic performance, underscoring the intricacy of accurate prediction in cybersecurity. Simpler models, including Logistic Regression, Linear Discriminant Analysis, and Gaussian Naive Bayes, demonstrated moderate to low accuracy, approximately around 70%. These findings stress the imperative need for a nuanced approach in the selection and fine-tuning of machine learning models to ensure effective DDoS detection in SDN environments.&lt;/div&gt;

Executive functioning is associated to everyday interference of pain in patients with chronic pain.

Dysfunction in executive functions is common among patients with chronic pain. However, the relationships between executive functioning and pain management have not been extensively studied. In this study, 189 outpatients (160 women, 29 men; mean age 33.15) with chronic pain underwent an extensive neuropsychological assessment, including several tests of executive functions. In addition, all participants completed self-assessment questionnaires regarding pain and interference of pain in everyday life. After adjusting for effects of age, education, and depression, several aspects of executive functioning were significantly associated with self-assessed everyday interference of pain (rs = 0.13-0.22, all ps < 0.05). This indicates that lower performance on tests of executive functioning was significantly associated with a higher degree of pain interference and a lower degree of life control. Pain characteristics such as pain intensity, pain duration, and pain spreading were not associated with executive functioning. These results suggest that preserved executive functions are related to better coping with pain, but not directly to the pain itself, in patients with chronic pain. Depression was also associated with self-management of pain, indicating that patients with lower executive functioning in combination with depression may need special attention during rehabilitation.

Early detection of verticillium wilt in eggplant leaves by fusing five image channels: a deep learning approach.

As one of the world's most important vegetable crops, eggplant production is often severely affected by verticillium wilt, leading to significant declines in yield and quality. Traditional multispectral disease-imaging equipment is expensive and complicated to operate. Low-cost multispectral devices cannot capture images and cover less information. The traditional approach to early disease diagnosis involves using multispectral disease-imaging equipment in conjunction with machine learning technology. However, this approach has significant limitations in early disease diagnosis, including challenges such as high costs, complex operation, and low model performance. The aim of this study was to combine low-cost multispectral cameras with deep learning technology to detect early stage Verticillium wilt in eggplant effectively. Using the Manual FS-3200T-10GE-NNC multispectral camera to perform multispectral imaging of the leaves of eggplant seedlings at the early infection stage, information fusion was performed on the collected multispectral images, and a five-channel image information fusion model was established. Image information fusion technology was combined with deep learning technology, among which the VGG16-triplet attention model performed the best, achieving a precision of 86.73% on the test set. Model validation on 48- and 72-hour data reached a precision of 75% and 82%, respectively, achieving an early diagnosis of Verticillium wilt. This highlighted the potential of multispectral cameras for early disease detection. In this study, we successfully developed a method for the non-destructive detection of the early stages of eggplant wilt disease by combining multispectral imaging technology with deep learning algorithms. While ensuring high accuracy, this method significantly reduces the cost of experimental equipment. The application of this method can reduce the cost of agricultural equipment and provide a scientific basis for agricultural production practices, helping to reduce losses caused by diseases.

Effects of Leadership Styles on Employees Job Performance: A Case Study of a Public Sector Organization in Ghana

Institutions are supervised by managers who are expected to direct their subordinates towards the achievements of organizational goals. Achievement of goals in an organization largely depends on the relationship between the supervisors and subordinates. Leadership styles play a crucial role in every organization in terms of industrial harmony and achievement of goals. Managers depending on the situation could employ democratic, autocratic or laissez-faire leadership styles to get organizational goals and objectives achieved. Primary data from public sector employees was used for this study based on the objectives of the study. Data was gathered from 50 employees using closed-ended questionnaires. The number of respondents was selected randomly from a population of 135 and determined using Yamane, 1974 formula and Browly proportional formula as indicated by Kumar 1976. The questionnaire were divided into three sections thus background characteristics of the respondents, respondents’ perception of their superiors’ leadership styles, employees’ assessment of their job performance respectively. The leadership styles were categorized autocratic, democratic and laissez-faire. The leadership and performance scales were measured using a 5-point Likert scale thus, 5=strongly agree; 4=agree; 3=undecided; 2=disagree; 1=strongly disagree. The outcome of the research was to the effect leadership exist in organizations have positive or negative impact on the job performance of workers. The results showed that a democratic leadership style has a significant positive impact on job performance within an organization, while autocratic and laissez-faire leadership styles contribute to apathy and lower work performance. Moreover, it was suggested among other things that the use of laissez-faire and autocratic styles should be discouraged since they are could be a recipe of creating lawless working environment which could result in low productivity.

A psychological reaction of higher education students to the new national exit exam in the case of Dilla University, 2023

IntroductionTest anxiety is an emotional state characterized by physiological and behavioral responses linked to the fear of poor exam results. It can result in a significant impact in the overall academic achievement of students. Addressing the contributing factors of the problem is better to intervene in the academic challenges of students and create a conducive learning environment. Thus, this study investigated the association between test anxiety and the newly implemented national exit exam program among prospective graduate students at Dilla University in 2023.MethodsA community-based cross-sectional design was employed, involving 357 participants recruited from August 15 to September 14, 2023. Data collection utilized structured questionnaires combined with interviews. The Westside Test Anxiety Inventory (WTAI) assessed test anxiety levels. Epi Info version 7 facilitated data entry, with subsequent analysis conducted using SPSS version 25. A p &amp;lt; 0.05 was used to determine statistical significance in the multivariate logistic regression model.ResultsThe prevalence of test anxiety associated with the national exit exam among prospective graduates at Dilla University was 68.9% (95% CI: 63.9–73.7%). The analysis identified significant associations between test anxiety and several factors: non-formal educational background of parents (AOR = 2.84, 95% CI: 2.12–4.65), moderate social support (AOR = 0.24, 95% CI: 0.12–0.46), and poor coping mechanisms (AOR = 3.67, 95% CI: 2.45–5.67).ConclusionThis study revealed a substantial prevalence of test anxiety among graduating students about the national exit exam. The findings suggest that factors such as limited social support and inadequate coping mechanisms contribute to test anxiety. Targeted interventions, including social support programs, counseling services, coping mechanism training, parental education, and support for students with lower academic performance, may be beneficial in mitigating test anxiety.

Immune-related adverse events in cancer patients referred to the palliative care team of a tertiary care center: a retrospective observational study.

The application of immune checkpoint inhibitors (ICIs) can cause multi-organ adverse events, namely immune-related adverse events (irAEs) in patients with cancer. This study aimed to characterize the epidemiological information on irAEs in patients with cancer referred to the palliative care team (PCT). The medical records of cancer patients with a history of ICI therapy referred to the PCT at a tertiary care center between January 2017 and July 2022 were retrospectively reviewed in this single-center, observational study. The median age of the 140 patients was 68years, and lung (39.3%) being the most common primary site. We observed irAEs in 46 patients (32.9%), and hypothyroidism was the most common irAE. For irAEs graded ≥ 3 in 21 patients, pneumonitis was the most common adverse event. As for strategies for management with irAEs, seventeen patients were treated with systemic steroids. irAEs ( +) had a significantly lower Performance Status at the start of ICI, a higher total number of ICI doses, and longer duration from start of ICI to date of death than irAEs (-). Among 140 cancer patients with a history of ICIs therapy consulted to the PCT, the prevalence of irAEs was 32.9%, and 21 patients (15.0%) developed irAEs with grade ≥ 3. As the use of ICI is expected to increase in the future, it is important for palliative care physicians to increase their awareness of the management of irAEs and collaborate with oncologists from an early stage.

Impact of Feature Selection Techniques on the Performance of Machine Learning Models for Depression Detection Using EEG Data

Major depressive disorder (MDD) poses a significant challenge in mental healthcare due to difficulties in accurate diagnosis and timely identification. This study explores the potential of machine learning models trained on EEG-based features for depression detection. Six models and six feature selection techniques were compared, highlighting the crucial role of feature selection in enhancing classifier performance. This study investigates the six feature selection methods: Elastic Net, Mutual Information (MI), Chi-Square, Forward Feature Selection with Stochastic Gradient Descent (FFS-SGD), Support Vector Machine-based Recursive Feature Elimination (SVM-RFE), and Minimal-Redundancy-Maximal-Relevance (mRMR). These methods were combined with six diverse classifiers: Logistic Regression, Support Vector Machine (SVM), Random Forest, Extreme Gradient Boosting (XGBoost), Categorical Boosting (CatBoost), and Light Gradient Boosting Machine (LightGBM). The results demonstrate the substantial impact of feature selection on model performance. SVM-RFE with SVM achieved the highest accuracy (93.54%) and F1 score (95.29%), followed by Logistic Regression with an accuracy of 92.86% and F1 score of 94.84%. Elastic Net also delivered strong results, with SVM and Logistic Regression both achieving 90.47% accuracy. Other feature selection methods yielded lower performance, emphasizing the importance of selecting appropriate feature selection and machine learning algorithms. These findings suggest that careful selection and application of feature selection techniques can significantly enhance the accuracy of EEG-based depression detection.

Low-Cost Laser Security System with Intrusion Detection and Alert Mechanism

Security becomes necessary factor nowadays. The crime gang improves their technology to perform their operation along with the development of technology. Hence, technology of security should be more improved to protect the crime works. We have decided to make a security project. In this project we have used laser light for security purpose. We know laser light goes through long distance. When any person or object passes through the laser line, the security alarm will start to ring. There are two parts of the system. One is transmitter and other is receiver. The transmitter side consists of a pair of dry cell batteries and an on-off switch. The receiver side, there is a focusing LDR (Light Dependent Resistor) sensor to sense the laser ray. It is connected with the main driver circuit which has two parts. One is signal of discontinuity ray and other is alarm circuit. When anybody passes through the ray, the main circuit sense the discontinuity and turn on the alarm circuit. The alarm will be ringing until the off button is pressed. There are two methods of ringing. One is the duration of ringing depends on preset timer and another is reset manually. Any option can be set by DPDT switch. We have used second method for our project. Mirrors are attached at every corner to reflect the ray coming from the laser. The reflected ray coming from the last mirror will hit the LDR directly. The system has built with low cost and high performance. The power consumption of the system is very low.

Similar adaptive responses in the upper body physical performance of table tennis players following the traditional and cluster set resistance and plyometric training.

Numerous studies have investigated the impact of resistance training (RT) and plyometric training (PT) set configurations (cluster sets [CS] and traditional sets [TS]) on lower body performance adaptations. However, the effects of these training modalities on upper body physical performance, particularly among Table Tennis (TT) players, remain unclear. Therefore, this study aimed to examine the effects of an 8-week RT and PT using CS and TS configurations on the physical performance attributes of male TT players. Forty young male players were recruited from a local academy and after determining their dominant hand, divided into CS-PT, TS-PT, CS-RT, TS-RT, and active control (CON) groups, each of eight. Pre- and post-interventions, the upper body muscular power (medicine ball throw), maximal strength (bench press), reaction time, and arm crank anaerobic power test were measured. Following the intervention, the CON group exhibited no changes; however, the CS-PT, TS-PT, CS-RT, and TS-RT groups displayed significant (p = 0.001) improvements in muscular power (effect size [ES] = 1.21 [large], 1.02 [medium], 0.81 [medium], 0.64 [medium]), strength (ES = 0.40 [small], 0.43 [small], 0.62 [medium], 0.63 [medium]), reaction time (ES = - 0.71 [medium], - 0.68 [medium], - 0.83 [medium], - 0.61 [medium]), peak power (ES = 1.83 [large], 1.66 [large], 0.90 [medium], 0.67 [medium]), and mean power (ES = 0.89 [medium], 0.89 [medium], 0.57 [small], 0.59 [small]), respectively. The PT groups indicated significantly greater changes than the RT groups in both muscular (CS, p = 0.021; TS, p = 0.018) and anaerobic power (peak power: CS, p = 0.017; TS, p = 0.011; mean power: CS, p = 0.024; TS, p = 0.042) performance, while the RT groups showed greater adaptive responses than the PT groups in maximal strength (CS, p = 0.033; TS, p = 0.048). In conclusion, PT is recommended for enhancing power performance, while RT is preferred for increasing strength. When various set configurations were incorporated, no further effects were noted.

Performance of complementor’s complementary goods on platform ecosystems: the perspective of individual complementors' complementary goods development behavior

AbstractThe purpose of our study was to examine how individual complementors' development behaviors (specifically, diversity- and distinctiveness-oriented development behaviors) impact the performance of complementary goods on the platform. To achieve this, we employed fixed-effects regression analysis and utilized panel data obtained from the free music platform “DOVA.” This study examines two aspects of the performance of complementor’s complementary goods: the highest performance related to the success of the platform ecosystem and the lowest performance related to platform ecosystem sustainment. The findings indicate that a high focus on diversity-oriented development behavior is positively correlated with the complementor's highest performance and negatively correlated with complementor's lowest performance. Additionally, distinctiveness-oriented development behavior also positively correlated with complementor's highest performance. However, low distinctiveness-oriented development behavior is negatively correlated with complementor's lowest performance. We reveal desirable development behaviors for complementary goods related to both content innovation and platform ecosystem sustainability. Based on these understandings, platform owners can target platform governance.

Tl2XY (X, Y = S, Se) monolayers with low lattice thermal conductivity and high thermoelectric performance by full-band approach with four scattering mechanisms

Abstract The low lattice thermal conductivity and high thermoelectric performance of the Janus Tl2SSe monolayer in the temperature region of 300–700 K are identified based on the thermoelectric properties of the Tl2S2, Tl2SSe, and Tl2Se2 monolayers. The transport coefficients for carrier concentrations and temperatures are obtained by solving the linearized Boltzmann transport equation in a full-band electronic structure. Four scattering mechanisms of acoustic deformation potential, optical deformation potential, polar optical phonon, and ionized impurity scatterings are considered. The ionized impurity scattering is recognized as the most important. The lattice thermal conductivity of the Janus Tl2SSe monolayer is substantially smaller than those of the Tl2S2, Tl2Se2 monolayers with higher symmetry. Moreover, we find that the Janus structures of the Tl2SSe monolayer increase the dielectronic constants and enhance the polar optical phonon scattering, then reduce the power factor to some extent. Therefore, the lattice thermal conductivity actually couples with the transport coefficient and cannot be individually regulated as is usually assumed. However, the ZT value of the Tl2SSe monolayer can still reach 1.77 at 700 K even if the intrinsic concentration and the bipolar effect are included. Therefore, the Tl2SSe monolayer is expected to be a promising candidate for thermoelectric materials.

Optimization of High-Speed Rail Express Transportation Scheduling Based on Spatio-Temporal-State Network Modeling

High-speed rail express is gradually gaining attention due to its low cost and high on-time performance. This paper studied the optimization of high-speed rail express transportation modes considering passenger capacity loss. A spatio-temporal network was constructed that includes the fixed passenger train schedule and the high-speed rail express transportation status. Based on this network, an optimization model for transportation schemes considering passenger capacity loss with non-splittable freight flows was developed. The model aims to minimize passenger loss and maximize net revenue from express transportation while comprehensively considering practical constraints such as transfer, loading and unloading, and carrying capacity. The Gurobi optimization solver was used to solve numerical cases, and the optimization results and solution effectiveness were comparatively analyzed under scenarios of sufficient and insufficient capacity.

Hybrid Oversampling and Undersampling Method (HOUM) via Safe-Level SMOTE and Support Vector Machine

The improvements in collecting and processing data using machine learning algorithms have increased the interest in data mining. This trend has led to the development of real-life decision support systems (DSSs) in diverse areas such as biomedical informatics, fraud detection, natural language processing, face recognition, autonomous vehicles, image processing, and each part of the real production environment. The imbalanced datasets in some of these studies, which result in low performance measures, have highlighted the need for additional efforts to address this issue. The proposed method (HOUM) is used to address the issue of imbalanced datasets for classification problems in this study. The aim of the model is to prevent the overfitting problem caused by oversampling and valuable data loss caused by undersampling in imbalanced data and obtain successful classification results. The HOUM is a hybrid approach that tackles imbalanced class distribution challenges, refines datasets, and improves model robustness. In the first step, majority-class data points that are distant from the decision boundary obtained via SVM are reduced. If the data are not balanced, SLS is employed to augment the minority-class data. This loop continues until the dataset becomes balanced. The main contribution of the proposed method is reproducing informative minority data using SLS and diminishing non-informative majority data using the SVM before applying classification techniques. Firstly, the efficiency of the proposed method, the HOUM, is verified by comparison with the SMOTE, SMOTEENN, and SMOTETomek techniques using eight datasets. Then, the results of the W-SIMO and RusAda algorithms, which were developed for imbalanced datasets, are compared with those of the HOUM. The strength of the HOUM is revealed through this comparison. The proposed HOUM algorithm utilizes a real dataset obtained from a project endorsed by The Scientific and Technical Research Council of Turkey. The collected data include quality control and processing parameters of yarn data. The aim of this project is to prevent yarn breakage errors during the weaving process on looms. This study introduces a decision support system (DSS) designed to prevent yarn breakage during fabric weaving. The high performance of the algorithm may encourage producers to manage yarn flow and enhance the HOUM’s efficiency as a DSS.

Performance of a Deep Learning System and Performance of Optometrists for the Detection of Glaucomatous Optic Neuropathy Using Colour Retinal Photographs

Background/Objectives: Glaucoma is the leading cause of irreversible blindness, with a significant proportion of cases remaining undiagnosed globally. The interpretation of optic disc and retinal nerve fibre layer images poses challenges for optometrists and ophthalmologists, often leading to misdiagnosis. AI has the potential to improve diagnosis. This study aims to validate an AI system (a convolutional neural network based on the Inception-v3 architecture) for detecting glaucomatous optic neuropathy (GON) using colour fundus photographs from a UK population and to compare its performance against Australian optometrists. Methods: A retrospective external validation study was conducted, comparing AI’s performance with that of 11 AHPRA-registered optometrists in Australia on colour retinal photographs, evaluated against a reference (gold) standard established by a panel of glaucoma specialists. Statistical analyses were performed using sensitivity, specificity, and area under the receiver operating characteristic curve (AUROC). Results: For referable GON, the sensitivity of the AI (33.3% [95%CI: 32.4–34.3) was significantly lower than that of optometrists (65.1% [95%CI: 64.1–66.0]), p &lt; 0.0001, although with significantly higher specificity (AI: 97.4% [95%CI: 97.0–97.7]; optometrists: 85.5% [95%CI: 84.8–86.2], p &lt; 0.0001). The optometrists demonstrated significantly higher AUROC (0.753 [95%CI: 0.744–0.762]) compared to AI (0.654 [95%CI: 0.645–0.662], p &lt; 0.0001). Conclusion: The AI system exhibited lower performance than optometrists in detecting referable glaucoma. Our findings suggest that while AI can serve as a screening tool, both AI and optometrists have suboptimal performance for the nuanced diagnosis of glaucoma using fundus photographs alone. Enhanced training with diverse populations for AI is essential for improving GON detection and addressing the significant challenge of undiagnosed cases.

Pre-Impact Fall Detection for E-Scooter Riding Using an IMU: Threshold-Based, Supervised, and Unsupervised Approaches

Pre-impact fall detection during e-scooter riding is essential for rider safety. Both threshold-based and deep learning algorithms (supervised and unsupervised models) were developed in this study. Twenty participants performed normal driving maneuvers such as straight driving, speed bumps, clockwise roundabouts, and counterclockwise roundabouts, along with falls (abnormal driving maneuvers). A 6-axis IMU sensor (Xsens DOT, The Netherlands) was positioned at the T7 location to record data at 60 Hz. The approaches included threshold-based, supervised learning, and unsupervised learning models The threshold-based approach yielded an accuracy of 98.86% with an F1 score of 0.99, while the supervised model had a slightly lower performance, reaching 86.29% accuracy and an F1 score of 0.56. The unsupervised knowledge distillation model achieved 98.86% accuracy, an F1 score of 0.99, and a memory size of only 46 kB. All models demonstrated lead times of more than 250 ms, sufficient for airbag deployment.