Search Results (164)

This study applies Gradient Boosting Machines (GBMs) and principal component regression (PCR) to forecast the closing price of the Johannesburg Stock Exchange (JSE) All-Share Index (ALSI), using daily data from 2009 to 2024, sourced from the Wall Street Journal. The models are evaluated under three training–testing split ratios to assess short-term forecasting performance. Forecast accuracy is assessed using standard error metrics: mean absolute error (MAE), root mean square error (RMSE), mean absolute percentage error (MAPE), and mean absolute scaled error (MASE). Across all test splits, the GBM consistently achieves lower forecast errors than PCR, demonstrating superior predictive accuracy. To validate the significance of this performance difference, the Diebold–Mariano (DM) test is applied, confirming that the forecast errors from the GBM are statistically significantly lower than those of PCR at conventional significance levels. These findings highlight the GBM’s strength in capturing nonlinear relationships and complex interactions in financial time series, particularly when using features such as the USD/ZAR exchange rate, oil, platinum, and gold prices, the S&P 500 index, and calendar-based variables like month and day. Future research should consider integrating additional macroeconomic indicators and exploring alternative or hybrid forecasting models to improve robustness and generalisability across different market conditions. Full article

Background: Marker-based motion capture is the current gold standard for three-dimensional (3D) gait analysis. This is a highly technical analysis that is time-consuming, and marker application can trigger anxiety in children. One potential solution is to use markerless camera systems instead. The objective of this study was to compare 3D lower limb gait kinematics in children using both marker-based and markerless motion capture methods. Methods: Ten typically developing children (age 6–13 yrs) completed five barefoot walks at a self-selected speed. A 10-camera marker-based system (Oqus, Qualisys) and a 7-camera markerless system (Miqus, Qualisys) captured synchronised gait data at 85 Hz. Generalised Additive Mixed Models were fitted to the data to identify the random effects of measurement systems, age, and time across the gait cycle. The root-mean-square difference (RMSD) was used to compare the differences between systems. Results: Significant interactions and differences were observed between the marker-based and markerless systems for most joint angles and planes of motion, particularly with regard to time and age. Conclusions: Despite differences across all kinematic profiles, the RMSD in this study was comparable to previously published results. Alternative model definitions and kinematic crosstalk in both systems likely explain the differences. Age differences were not consistent across joint levels, suggesting a larger sample size is required to determine how maturation may affect markerless tracking. Further investigation is required to understand the deviations and differences between systems before implementing markerless technology in a clinical setting. Full article

This paper addresses the critical need for detailed electricity and peak power demand maps for urban public transportation vehicles. Current approaches often rely on overly general assumptions, leading to considerable errors in specific applications or, conversely, overly specific measurements that limit generalisability. We aim to present a comprehensive data-driven methodology for analysing energy consumption within a large urban agglomeration. The method leverages a unique and extensive set of real-world performance data, collected over two years from onboard recorders on all public bus lines in the Capital City of Warsaw. This large dataset enables a robust probabilistic analysis, ensuring high accuracy of the results. For this study, three representative bus lines were selected. The approach involves isolating inter-stop trips, for which instantaneous power waveforms and energy consumption are determined using classical mathematical models of vehicle drive systems. The extracted data for these sections is then characterised using probability distributions. This methodology provides accurate calculation results for specific operating conditions and allows for generalisation with additional factors like air conditioning or heating. The direct result of this paper is a detailed urban map of energy demand and peak power for public transport vehicles. Such a map is invaluable for planning new traffic routes, verifying existing ones regarding energy consumption, and providing a reliable input source for strategic charger deployment analysis along the route. Full article

With the goal of addressing the difficulty of rapidly identifying the source of fire in commercial buildings, this study builds a numerical fire model based on the fire dynamics simulator (FDS) and combines it with a hybrid model to predict the location of a fire source. Different scenarios were built to simulate the spatial and temporal distributions of key parameters such as temperature, smoke, and CO concentration during the fire process. Combining convolutional neural networks (CNNs) and support vector machines (SVMs) for prediction, the fire-source location prediction model with temperature, smoke, and CO concentration as feature quantities was constructed, and the hyperparameters affecting the model accuracy and generalisation were optimised by the Crested Porcupine Optimizer (CPO) algorithm. The experimental results show that the positioning error of this method under the building plane is less than 0.95 m, the mean absolute error (MAE) is within 0.35, and the root-mean-square error (RMSE) is within 0.41, which are 43% and 82% higher than the unoptimised model, respectively. The localisation accuracy of the fire-source room is 97.61%. In addition, the model’s anti-interference performance was tested under various extreme conditions. The results show that the proposed model can ensure the accurate location of a fire source and can provide information in emergencies. Full article

It is shown that the extensive experimental material available in the literature reflecting the behaviour of thermosensitive hydrogels and solutions of thermosensitive polymers requires systematisation and generalisation. Additional evidence is given that the method of forward and reverse phase portraits is an important tool for the systematisation of such data. It is shown that using this method makes it possible to refine the characteristics of the phase transition, as well as to classify thermosensitive hydrogels and solutions according to such classification criteria as the number of phase transition stages. Based on the developed classification, a new semi-empirical theory of phase transitions is proposed. Using this model, it is shown for the first time that phase transitions can be described through equivalent chemical reactions of the first and second orders. The proposed model allows us to explain the fact that the phase portraits obtained from experimental data often contain segments corresponding to parabolic and linear dependences. It is shown that the proposed approach creates a basis for systematisation of the results accumulated in the field of study of thermosensitive polymers in automatic mode by means of image recognition technologies. Full article

The study of personalities is a major component of human psychology, and with an understanding of personality traits, practical applications can be used in various domains, such as mental health care, predicting job performance, and optimising marketing strategies. This study explores the prediction of Big Five personality trait scores from online comments using transformer-based language models, focusing on improving the model performance with a larger dataset and investigating the role of intercorrelations among traits. Using the PANDORA dataset from Reddit, the RoBERTa and BERT models, including both the base and large variants, were fine-tuned and evaluated to determine their effectiveness in personality trait prediction. Compared to previous work, our study utilises a significantly larger dataset to enhance the model’s generalisation and robustness. The results indicate that RoBERTa outperforms BERT across most metrics, with RoBERTa large achieving the best overall performance. In addition to evaluating the overall predictive accuracy, this study investigates the impact of intercorrelations among personality traits. A comparative analysis is conducted between a single-model approach, which predicts all five traits simultaneously, and a multiple-model approach, fine-tuning the models independently and each predicting a single trait. The findings reveal that the single-model approach achieves a lower RMSE and higher $R^2$ values, highlighting the importance of incorporating trait intercorrelations in improving the prediction accuracy. Furthermore, RoBERTa large demonstrated a stronger ability to capture these intercorrelations compared to previous studies. These findings emphasise the potential of transformer-based models in personality computing and underscore the importance of leveraging both larger datasets and intercorrelations to enhance predictive performance. Full article

Traditional pest and disease identification methods mainly rely on manual detection or traditional machine learning techniques, but they have obvious deficiencies in terms of their accuracy and generalisation ability. In recent years, deep learning has gradually become the preferred solution for the intelligent identification of pests and diseases by virtue of its powerful automatic feature extraction and complex data-processing capabilities. In this paper, we systematically present the application of traditional machine learning methods in pest and disease identification and their limitations, and focus on the research progress of deep learning methods, covering three mainstream architectures: convolutional neural network (CNN), Vision Transformer model and CNN–Transformer hybrid model. In addition, this paper provides an in-depth analysis of the key challenges currently faced in the field of pest recognition, including the problems of small-sample learning, complex background interference and model lightweighting, and further propose solutions for future research to provide theoretical references and technical guidance for the development of related fields. Full article

This study investigates the effects of warming water on growth in seven demersal fish species including Atlantic cod (Gadus morhua), American plaice (Hippoglossoides platessoides), Greenland halibut (Reinhardtius hippoglossoides), roughhead grenadier (Macrourus berglax) and three species of redfish (Sebastes spp.) in the Northwest Atlantic and compares the changes in growth across species. Length-at-age data were collected from EU bottom trawl surveys from 1993 to 2018, and bottom temperature data were obtained from the Copernicus Marine Service. Generalised additive mixed models (GAMMs) were used to describe the temperature effects on growth. The analysis was carried out separately for males and females. Both sexes of all species except American plaice showed significant temperature effects on growth. To obtain the growth parameters, von Bertalanffy growth functions (VBGFs) were fitted to the predictions from best-fit GAMMs for all species and both sexes under five different bottom temperature scenarios (3, 3.5, 4, 4.5 and 5 °C). The predictions from all best-fit GAMMs were broadly similar in form to the fitted von Bertalanffy growth functions (R² > 90%). Increased bottom temperature generally resulted in a decrease in the asymptotic length (L_∞) and an increase in the growth rate (k). The species with the most dramatic increase in k over the temperature range of 3 °C to 5 °C was Atlantic cod, for which k increased from 0.05 to 0.13 year⁻¹ in females and from 0.08 to 0.14 year⁻¹ in males. The maximum length (L_max), predicted by the VBGF at maximum age generally declined from 3 °C to 5 °C. The species with the most pronounced decline in L_max was beaked redfish (S. mentella). An increase in the proportion of smaller individuals could impact population productivity and result in lower biomass available to fisheries. Uneven changes in fish growth in the warming ocean could also have wider ecological implications and alter the trophic landscape. Full article

Background: An unhealthy diet is a major risk factor for colorectal cancer (CRC). This study assessed the diet-related CRC burden from 1990 to 2021 in Organisation for Economic Co-operation and Development (OECD) nations and estimated the burden until 2050. Methods: Data for OECD countries on diet-related CRC disability-adjusted life years (DALYs) and deaths were obtained from the Global Burden of Disease 2021 study. The estimated annual percent change (EAPC) was calculated to analyse the CRC burden attributable to dietary factors. A generalised additive model with a negative binomial distribution was used to predict the future burden of CRC attributable to dietary factors from 2021 to 2050. Results: In 2021, the age-standardised percentages of diet-related CRC DALYs and deaths were 39.1% (95% uncertainty interval (UI): 9.3, 61.3) and 39.0% (95% UI: 9.7, 60.9), respectively, in the OECD countries. Between 1990 and 2021, the age-standardised DALYs decreased from 185 to 129 per 100,000, and deaths decreased from 8 to 6 per 100,000 population for OECD countries. Similarly, the EAPC in the rates showed a downward trend (EAPC_deaths = −1.26 and EAPC_DALYs = −1.20). The estimated diet-related CRC DALYs and deaths are projected to increase to 4.1 million DALYs and 0.2 million deaths by 2050. There is a downward trend in CRC deaths (EAPC = 1.33 for both sexes) and in DALYs (−0.90 for males and −1.0 for females) from 1990 to 2050. Conclusions: The diet-related CRC burden remains significant. Implementing nutrition intervention programmes is necessary to promote access to affordable and nutritious foods and raise awareness about the importance of a healthy diet in reducing CRC risk. Full article

In this work, we establish the existence of positive solutions for a problem driven by a multi-phase operator composed of two distinct exponent Laplacian-type operators and a generalised lower-order term, which ensures asymmetric behaviour across three subregions of the domain under consideration. The reaction term involves a mild singularity at zero and includes a possibly sign-changing perturbation function. Under additional restrictive conditions, we also obtain a uniqueness result for the problem. Our existence result is based on pseudomonotone operator theory. Moreover, a detailed analysis, combined with a Díaz–Saá-type argument, allows us to also establish a uniqueness theorem. To the best of our knowledge, this is the first work addressing such a generalisation of the multi-phase operator. These novel results can serve as a foundation for more general physical and engineering models. Full article

This study proposes a novel modelling approach that integrates mutual information (MI)-based parameter screening with random forest (RF) modelling to achieve an accurate quantitative prediction of surface hardness and impact energy in two martensitic stainless steels (1Cr13 and 2Cr13). Preliminary analyses indicated that the magnetic parameters derived from Barkhausen noise (MBN), and the incremental permeability (IP) measurements showed limited linear correlations with the target properties (surface hardness and impact energy). To address this challenge, an MI feature screening method has been developed to identify both the linear and non-linear parameter dependencies that are critical for predicting target mechanical properties. The selected features were then fed into an RF model, which outperformed traditional multiple linear regression in handling the complex, non-monotonic relationships between magnetic signatures and mechanical performance. A key advantage of the proposed MI-RF framework lies in its robustness to small sample sizes, where it achieved high prediction accuracy (e.g., R² > 0.97 for hardness, and R² > 0.86 for impact energy) using limited experimental data. By leveraging MI’s ability to capture multivariate dependencies and RF’s ensemble learning power, it effectively mitigates overfitting and improves generalisation. In addition to demonstrating a promising tool for the non-destructive evaluation of martensitic steels, this study also provides a transferable paradigm for the quantitative assessment of other mechanical properties by magnetic feature fusion. Full article

Small defects on the surface of copper strips have a significant impact on key properties such as electrical conductivity and corrosion resistance, and existing inspection techniques struggle to meet the demand in terms of accuracy and generalisability. Although there have been some studies on metal surface defect detection, there is a relative lack of research on highly reflective copper strips. In this paper, a lightweight and efficient copper strip defect detection algorithm, SC-AttentiveNet, is proposed, aiming to solve the problems of the large model size, slow speed, insufficient accuracy and poor generalisability of existing models. The algorithm is based on ConvNeXt V2, and combines the SCDown module and group normalisation to design the SCGNNet feature extraction network, which significantly reduces the computational overhead while maintaining excellent feature extraction capability. In addition, the algorithm introduces the SPPF-PSA module to enhance the multi-scale feature extraction capability, and constructs a new neck feature fusion network via the HD-CF Fusion Block module, which further enhances the feature diversity and fine granularity. The experimental results show that SC-AttentiveNet has a mAP of 90.11% and 64.14% on the KUST-DET and VOC datasets, respectively, with a parameter count of only 6.365 MB and a computational complexity of 14.442 GFLOPs. Tests on the NEU-DET dataset show that the algorithm has an excellent generalisation performance, with a mAP of 76.41% and a detection speed of 78 FPS, demonstrating a wide range of practical application potential. Full article

Due to the widespread applicability of generative artificial intelligence, we have seen it adopted across many areas of education, providing universities with new opportunities, particularly in cybersecurity education. With the industry facing a skills shortage, this paper explores the use of generative artificial intelligence in higher cybersecurity education as an intelligent tutoring system to enhance factors leading to positive student outcomes. Despite its success in content generation and assessment within cybersecurity, the field’s multidisciplinary nature presents additional challenges to scalability and generalisability. We propose a solution using agents to orchestrate specialised large language models and to demonstrate its applicability in graduate level cybersecurity topics offered at a leading Australian university. We aim to show a generalisable and scalable solution to diversified educational paradigms, highlighting its relevant features, and a method to evaluate the quality of content as well as the general effectiveness of the intelligent tutoring system on subjective factors aligned with positive student outcomes. We further explore areas for future research in model efficiency, privacy, security, and scalability. Full article

Cardiovascular diseases are the most common cause of death worldwide, with atmospheric pollution, and primarily particulate matter, standing out as the most hazardous environmental factor. To explore the exposure–response curves, traditional epidemiological studies rely on generalised additive or linear models and numerous works have demonstrated the relative risk and the attributable fraction of mortality/morbidity associated with exposure to increased levels of particulate matter. An alternative, probably more effective, procedure to address the above issue is using machine learning models, which are flexible and often outperform traditional methods due to their ability to handle both structured and unstructured data, as well as having the capacity to capture non-linear, complex associations and interactions between multiple variables. This study uses five advanced machine learning techniques to examine the contribution of several climatic factors and air pollutants to cardiovascular mortality in the Eastern Mediterranean region, focusing on Thessaloniki, Greece, and Limassol, Cyprus, covering the periods 1999–2016 and 2005–2019, respectively. Our findings highlight that temperature fluctuations and major air pollutants significantly affect cardiovascular mortality and confirm the higher health impact of temperature and finer particles. The lag analysis performed suggests a delayed effect of temperature and air pollution, showing a temporal delay in health effects following exposure to air pollution and climatic fluctuations, while the seasonal analysis suggests that environmental factors may explain greater variability in cardiovascular mortality during the warm season. Overall, it was concluded that both air quality improvements and adaptive measures to temperature extremes are critical for mitigating cardiovascular risks in the Eastern Mediterranean. Full article

With growing worldwide interest in constructing larger and taller wooden buildings, wood properties, such as the dynamic modulus of elasticity (${MOE}_{dyn}$), have become increasingly important. However, the ${MOE}_{dyn}$ of trees and logs has rarely been considered in forest management because a method for estimating the ${MOE}_{dyn}$ of logs based on standing tree characteristics has been lacking. Herein, we explored the multiple relationships between the ${MOE}_{dyn}$ of logs and standing tree characteristics of Japanese cedar (Cryptomeria japonica) such as tree height, diameter at breast height (DBH), and tree age, including the stress-wave velocity of the tree, which is known to be correlated with the ${MOE}_{dyn}$ of logs. The relationship between the ${MOE}_{dyn}$ of logs and standing tree characteristics was investigated by considering the bucking position. Different trends between the bottom logs and upper logs were found for all characteristics, showing a multiple trend of tree characteristics with the ${MOE}_{dyn}$ of logs based on the bucking position. The top three generalised linear mixed models for the prediction of the ${MOE}_{dyn}$ of logs showed relatively high accuracies when the bucking position was considered as a random effect. Although the contribution of the stress-wave velocity of the tree was relatively high, adding tree age improved the accuracy of the model, and this model was selected as the top model. The model for the bottom log, utilising the stress-wave velocity and age of the tree as explanatory variables, was highly explanatory (R² = 0.70); however, the best model for upper logs was only moderately explanatory (R² = 0.44). In addition, tree height and DBH were selected as explanatory variables along with tree age in the second and third models, which suggested the importance of growth rate rather than tree size. Therefore, adding correlates associated to characteristics related to height growth, such as site index, and DBH growth, such as stand density, is expected to improve model accuracy. Full article