Search Results (1,628)

The fundamental goal of a construction project is to complete the construction phase within budget, but in practice, planned cost estimates are often exceeded. The causes of overruns can be due to insufficient preparation and planning of the project, changes during construction, activation of risky events, etc. Also, construction costs are often calculated based on experience rather than scientifically based approaches. Due to the challenges, this paper investigates the potential of several different machine learning methods (linear regression, decision tree forest, support vector machine and general regression neural network) for estimating construction costs. The methods were implemented on a database of recent high-rise construction projects in the Republic of Croatia. Results confirmed the potential of the selected assessment methods; in particular, the support vector machine stands out in terms of accuracy metrics. Established machine learning models contribute to a deeper understanding of real construction costs, their optimization, and more effective cost management during the construction phase. Full article

Volatility in the modern world and electricity Day-Ahead Markets (DAMs) usually makes long-term historical data irrelevant or even detrimental for accurate forecasting. This study directly addresses this challenge by proposing a novel forecasting paradigm centered on extremely short training windows, ranging from 7 to 90 days, to maximize responsiveness to recent market dynamics. This volatility-driven approach intentionally creates a data-scarce environment where the suitability of deep learning models is limited. Building on the hypothesis that shallow machine learning models, and more specifically boosting trees, are better adapted to this reality, we evaluate four models, namely LSTM with feed-forward error correction, XGBoost, LightGBM, and CatBoost, across three European energy markets (Greece, Belgium, Ireland) using feature sets derived from ENTSO-E forecast data. Results consistently demonstrate that LightGBM provides superior forecasting accuracy and robustness, particularly when trained on 45–60 day windows, which strike an optimal balance between temporal relevance and learning depth. Furthermore, a stronger capability in detecting seasonal effects and peak price events is exhibited. These findings validate that a short-window training strategy, combined with computationally efficient shallow models, is a highly effective and practical approach for navigating the volatility and data constraints of modern DAM forecasting. Full article

Hydrological behavior plays a critical role in seasonally dry forest ecosystems, as it underpins water availability for multiple productive activities, including forestry, agriculture, grazing, and urban supply. This study evaluated the hydrological effects of thinning treatments in a semi-arid oak–pine forest of Chihuahua, Mexico, using a Before–After–Control–Impact (BACI) design. Three Micro-catchments (MC) with initially comparable tree density and canopy cover were monitored during the rainy seasons of 2018 (pre-thinning) and 2019 (post-thinning). Thinning treatments were applied at 20% and 60% canopy cover in two MC, while a third remained unthinned as a 100% control. Precipitation and surface runoff were recorded at the event scale, and data were analyzed using Weibull probability models with a log link to capture the frequency and magnitude of runoff events. Precipitation patterns were broadly comparable across years, although 2018 included an extreme storm event (59 mm). In contrast, runoff volumes in 2019 were lower despite marginally higher seasonal rainfall, reflecting the absence of large storms. Statistical modeling indicated that for each additional millimeter of precipitation, mean runoff increased by approximately 12%, although thinning significantly altered baseline conditions. Relative to 2018, mean runoff ratios were 0.087 in the 100% canopy catchment, 0.296 in the 60% treatment, and 0.348 in the 20% treatment, suggesting that reduced canopy cover retained proportionally more runoff than the control. BACI contrasts confirmed that thinned catchments maintained higher proportions of runoff than the unthinned control, although statistical significance was marginal for the 20% canopy treatment. Overall, the study provides ecohydrological insights relevant to the management of semi-arid forest ecosystems. Full article

LoRaWAN is widely used for IoT environmental monitoring, but its lightweight security mechanisms leave the physical layer vulnerable to availability attacks such as jamming and battery-depletion. These risks are particularly critical in mission-critical environmental monitoring systems. This paper proposes a multi-attribute physical-layer authentication (PLA) framework that supports uplink legitimacy assessment by jointly exploiting radio, energy, and temporal attributes, specifically RSSI, altitude, battery_level, battery_drop_speed, event_step, and time_rank. Using publicly available Brno LoRaWAN traces, we construct a device-aware semi-synthetic dataset comprising 230,296 records from 1921 devices over 13.68 days, augmented with energy, spatial, and temporal attributes and injected with controlled jamming and battery-depletion anomalies. Five classifiers (Random Forest, Multi-Layer Perceptron, XGBoost, Logistic Regression, and K-Nearest Neighbours) are evaluated using accuracy, precision, recall, F1-score, and AUC-ROC. The Multi-Layer Perceptron achieves the strongest detection performance (F1-score = 0.8260, AUC-ROC = 0.8953), with Random Forest performing comparably. Deployment-oriented computational profiling shows that lightweight models such as Logistic Regression and the MLP achieve near-instantaneous prediction latency (below 2 µs per sample) with minimal CPU overhead, while tree-based models incur higher training and storage costs but remain feasible for Network Server-side deployment. Full article

Streptococcus pneumoniae (S. pneumoniae) is responsible for a wide range of infections. The aim of this study was to investigate the clonal diversity of S. pneumoniae in thirteen Arab countries. Multi-Locus Sequence Typing (MLST) data were extracted from PubMLST database. Genetic analysis was performed using DnaSP software version 6.0. A Minimum Spanning Tree (MST) analysis was conducted to evaluate the population structure of S. pneumoniae strains. Genetic data from 1008 Arab S. pneumoniae strains, collected over 22 years (1996–2018), were analyzed. MLST analysis identified a highly diverse population comprising 600 sequence types grouped into 87 clonal complexes and 295 singletons. Both internationally disseminated clones (e.g., ST156) and country-specific lineages (e.g., ST2307, Saudi Arabia) were observed, indicating substantial geographic structuring. Significant associations were detected between sequence types and geographical origin, decade of isolation, patient age, disease type, and serotype (p < 0.05). Although recombination events were presented, the population retained a predominantly clonal structure over time (I^S_A = 0.0715, p < 0.001). Overall, these findings demonstrated extensive genetic heterogeneity and spatiotemporal structuring of S. pneumoniae in the Arab region, providing valuable insights for regional surveillance and vaccine-related strategies. Full article

Ormosia henryi, a rare and endemic timber tree in China, possesses exceptional economic and ecological value, but it has experienced a critical decline in wild populations. We integrated PacBio HiFi and Hi-C technologies to generate a superior, chromosome-level genome assembly, establishing a more robust genetic foundation than existing draft sequences. The resulting assembly (2.64 Gb; Contig N50 = 39.17 Mb; and Scaffold N50 = 338.40 Mb) exhibits high continuity and completeness, effectively overcoming the assembly challenges associated with high heterozygosity (1.37%) and repetitive sequence content (83.89%). Comparative genomic analysis revealed that O. henryi diverged from Lupinus albus approximately 53.82 million years ago and underwent two independent whole-genome duplication events. The historical accumulation of evolutionary resilience is reflected in the significant expansion of 276 gene families enriched in photosynthesis and phenylpropanoid biosynthesis, alongside 122 genes under positive selection involved in DNA repair and proteostasis. These genomic signatures elucidate a stable genetic foundation. While wild populations have sharply declined in recent decades, this suggests that this status underscores the overwhelming impact of intense external anthropogenic pressures, such as overexploitation and habitat fragmentation, which may have overridden the species’ inherent adaptive capacity and slow life-history strategy. This high-quality genomic resource identifies key candidate loci, such as the PIF1 helicase for growth regulation, and provides a critical framework for screening elite germplasm for population restoration. Consequently, this study establishes a theoretical and molecular basis for transitioning from fundamental research to the precision conservation and sustainable industrial application of this high-value woody species. Full article

Bovine viral diarrhea virus (BVDV) primarily causes bovine viral diarrhea/mucosal disease, an infectious disease having a significant economic impact on the cattle-farming industry globally. Comprehensive monitoring and in-depth studies of the pathological characteristics of viruses are crucial in formulating effective prevention and control strategies. The isolation, identification, molecular characterization, and pathogenicity analysis of a BVDV strain isolated from persistently infected cattle ear tissue samples are reported in this study. This newly isolated strain is a noncytopathogenic BVDV, which we named HB2411. Homology between the HB2411 and U63479 strains was determined to be 96.7%, and the phylogenetic tree indicated that HB2411 belongs to the BVDV-1b subtype. Genetic variation analysis of the E2 protein of the HB2411 strain revealed multiple amino-acid mutation sites. Recombination analysis of the newly isolated HB2411 strain suggested a potential cross-geographical transmission event. BALB/c mice were intraperitoneally inoculated with the BVDV strain to evaluate the pathogenicity and virulence of BVDV-1b HB2411. BVDV was detected in multiple organs of BALB/c mice, with the highest viral load in the liver. BVDV infection promoted the expression of inflammatory cytokines in mice livers, necessitating further studies on the virulence and pathogenic mechanisms of this new strain to reduce economic losses caused to the animal husbandry industry. Full article

Hazard impacts are increasing in complexity as the world population grows. No universal strategies are available to minimize or eliminate the impacts of all scenarios. In this paper, a priority scenario-based strategy methodology is proposed using a Decision Tree (DT) machine learning tool. This approach identifies the parameters and combinations that contribute to high impact and loss from a hazard event conditioned on a priority scenario. The method is applied to a local water distribution network under seismic hazards. The priority scenarios in this study are vulnerability (VPS), damage (DPS), and cost (CPS). Each priority scenario identifies different affected areas. Some areas were repeatedly affected in different priority scenarios, showing an overlap of effects and making them a high crucial priority. Based on the analysis, a priority-based map was generated, highlighting areas that should be given priority for restoration or protection. The DTs were compared with other ML tools and Tree-based models to ascertain the best tool that determines the affected parameters. Competition tests compared the results from the ML tools and showed acceptable predictions; however, the DT was demonstrated to be the most ideal tool for this proposed method, showing an r² of 0.6745, 0.9259, and 0.7343 for VPS, DPS, and CPS, respectively. Full article

Quantifying forest patch dynamics is essential for understanding how forest patch characteristics vary in relation to ancient tree locations. This study developed a satellite-based framework to analyze the differences among forest patches associated with natural and planted ancient trees across the Sichuan–Chongqing region, China. Using monthly LandTrendr on Google Earth Engine, we analyzed long-term (1990–2024) and high-frequency observations of forest dynamics at a 180 m × 180 m (6 × 6 pixels) spatial scale. Disturbance and recovery events were characterized by their magnitude, rate, timing, and duration. Patches were classified into six categories based on ancient tree type and proximity and further subdivided by land use type. The results show that in natural forests, patches with natural ancient trees are associated with more stable change signatures, whereas in planted forests, patches containing planted ancient trees are associated with stronger recovery-related change patterns. Over 60% of detected changes were short-lived (≤5 years), indicating that most disturbances and recovery processes were transient rather than persistent. These findings show that the presence and spatial context of ancient trees are associated with differences in patch change patterns. The proposed workflow provides a scalable approach for integrating multi-temporal remote sensing into large-scale monitoring and management of ancient trees and their associated forest patches. Full article

Three organic soil amendments of different origins (chicken manure, fungal biomass obtained through biological fermentation, and a leonardite-based humic acid product) were applied to young chestnut trees, alongside mineral fertilizer, which when applied alone served as the control. During the second year, bud break pattern, photosynthetic activity, leaf carbohydrate concentrations, soil properties, and leaf nutrient content were evaluated across multiple sampling events. Sampling time significantly influenced most measured parameters. The addition of organic amendments accelerated bud break, influenced plant nutrient uptake, and modified soil properties. Notably, soil organic matter increased following chicken manure and fungal biomass applications, available phosphorus decreased under fungal biomass and leonardite-based humic acids (to 14.5 and 12.4 ppm, respectively, compared to 17.5 ppm in the mineral fertilizer control), and soil iron concentrations tripled under leonardite-based humic acids relative to the control. However, no significant effects were observed on photosynthetic performance or leaf carbohydrate concentrations. Discriminant and hierarchical cluster analyses revealed clear differences among amendments, with the humic acid-based product exerting distinct effects. As there are not many data available in the literature on the efficacy of organic amendments in chestnut cultivation, the present results underscore the importance of the site-specific selection of organic amendments, tailored to soil characteristics (in the present trial, an acidic soil) and specific nutritional objectives to optimize tree physiological performance. Full article

Under the background of global climate change, frequent drought events have significantly impacted plant–insect interaction. This study focuses on Ficus microcarpa, an important landscaping and urban greening tree species in tropical and subtropical regions, and its primary herbivorous pest, Perina nuda, by applying the age-stage, two-sex life table theory to systematically evaluate the effects on the life history traits and population dynamics of P. nuda reared on F. microcarpa subjected to different levels of drought stress. The results demonstrated that reared on drought-stressed F. microcarpa significantly altered multiple life history traits of P. nuda. All drought treatments significantly shortened the larval development period. Under both light and severe drought conditions, adult lifespan was prolonged, the total pre-oviposition period was reduced, fecundity per female increased, and generation time (T) was shortened. However, significant increases in pupal weight, intrinsic rate of increase (r), and finite rate of increase (λ) were observed only under light drought stress. The population prediction results indicate that both light and severe drought stress lead to obviously higher population growth rates and larger population sizes at 200 days compared to the control group. These findings suggest that the population fitness of P. nuda is enhanced under light and severe drought stress, potentially increasing the probability of pest outbreaks. This study provides an important theoretical basis and practical advice on forecasting population dynamics and implementing integrated management strategies for P. nuda in the context of climate change. Full article

Anomaly detection in financial transactions is a challenging task, primarily due to severe class imbalance and the adaptive behavior of fraudulent activities. This paper presents a reinforcement learning framework for fraud detection (RLFD) to address this problem. We train a deep Q-network (DQN) agent with a long short-term memory (LSTM) encoder to process sequences of financial events and identify anomalies. On a proprietary, highly imbalanced dataset, 10-fold cross-validation highlights a distinct trade-off in performance. While a gradient boosted trees (GBT) baseline demonstrates superior global ranking capabilities (higher ROC and PR AUC), the RLFD agent successfully learns a high-recall policy directly from the reward signal, meeting operational needs for rare event detection. Importantly, a dynamic orthogonality analysis proves that the two models detect distinct subsets of fraudulent activity. The RLFD agent consistently identifies unique fraudulent transactions that the tree-based model misses, regardless of the decision threshold. Even at high-confidence operating points, the RLFD agent accounts for nearly 30% of the detected anomalies. These results suggest that while tree-based models offer high precision for static patterns, RL-based agents capture sequential anomalies that are otherwise missed, supporting for a hybrid, parallel deployment strategy. Full article

Complications are frequent in cancer patients and contribute to adverse outcomes and higher healthcare costs, underscoring the need for earlier identification and prediction. This study evaluated the feasibility of using passively generated smartphone sensor data to explore early-warning signals of complications and symptom worsening during cancer treatment. A total of 108 patients were continuously monitored using accelerometer, GPS, and screen on/off data collected through the LAIMA application, while symptoms of depression, fatigue, and nausea were assessed every two weeks and complications were confirmed during clinic visits or emergency presentations. Smartphone data streams were aggregated into variables describing activity and sociability patterns. Machine learning models, including Decision Tree, Extreme Gradient Boosting, K-Nearest Neighbors, and Support Vector Machine, were used for complication prediction, and time-series models such as Autoregressive Integrated Moving Average, Holt–Winters, TBATS, Long Short-Term Memory neural network, and General Regression Neural Network were applied to identify early behavioral changes preceding symptom reports. In this exploratory analysis, the ensemble model demonstrated high sensitivity (89%) for identifying complication events. Smartphone-derived behavioral indicators enabled earlier detection of depression, fatigue, and vomiting by about nine days in a subset of patients. These findings demonstrate the feasibility of passive smartphone sensor data as exploratory early-warning signals, warranting validation in larger cohorts. Full article

Background and Objectives: The objective of this study is to assess the efficacy of a novel software risk score, PulsIn, in predicting cardiovascular diseases within an independent study conducted on subjects from the western region of Romania. Accurate prediction of cardiovascular events in hypertensive patients remains challenging when relying solely on traditional risk scores. This study proposes PulsIn, a composite risk score that integrates classical, echocardiographic, inflammatory, renal, and metabolic markers, combined with machine learning, to refine cardiovascular risk stratification. Materials and Methods: In a prospective cohort of 300 hypertensive adults without prior major cardiovascular events, we collected demographic and clinical data, standard risk factors, laboratory biomarkers (including homocysteine, paraoxonase-1 activity, microalbuminuria, and lipid profile), and advanced echocardiographic parameters (3D left ventricular ejection fraction, diastolic function, global longitudinal strain, and left atrial strain). PulsIn was constructed as an extended composite score and used as input to machine learning models (random forest, XGBoost, and other tree-based algorithms) to predict incident major cardiovascular events. Model performance was assessed by receiver operating characteristic curves, discrimination, calibration, and feature importance and compared with established risk scores (SCORE2, Framingham, QRISK, and others). Results: PulsIn-based models showed improved predictive performance compared with traditional scores, with XGBoost and random forest achieving area under the curve values up to approximately 0.85–0.88, versus 0.60–0.78 for conventional scores. Echocardiographic indices of subclinical cardiac damage, microalbuminuria, homocysteine, and paraoxonase-1 activity emerged as key predictors, particularly enhancing reclassification in patients at intermediate risk by traditional tools. Conclusions: The PulsIn composite risk score, integrating multimodal clinical, echocardiographic, and biomarker data within a machine learning framework, offers more accurate cardiovascular risk prediction than conventional algorithms in hypertensive patients. External validation in larger, independent, and more diverse populations is required before routine clinical implementation. Full article

Northeast British Columbia (54–60° N latitude, 120–123° W longitude) has 10+ M ha of complex conifer–broadleaf forest, which is a unique forest type in the province. Current management practice is to remove competing broadleaf species to promote the growth of more commercially valued conifers. This approach ignores the species silvics and results in forest simplification, thus reducing species and structural diversity, habitat value, and overall stand resilience to future events such as climate change and wildfires. These practices also negatively impact traditional First Nation treaty rights. Three trials were established across the region in 5-to-18-year-old post-logging mixed species stands where broadleaves had not been removed. Competition-free radii of 0, 1, 2, and 4 m were established around white spruce (Picea glauca (Moench) Voss) crop trees. The objective was to investigate the impact of broadleaf (aspen Populus tremuloides Michx. or paper birch Betula papyrifera Marsh.) competition on crop tree growth with respect to the free-to-grow (FTG) standard. Except at extreme broadleaf densities (>10,000 SPH), crop tree DBH growth was not impacted when trials were established. After at least 11 growing seasons, except at the competition-free 4 m radius, DBH was not impacted by competition. Spruce DBH in the mixed stand at all radii was greater than the expected BC model projections for a pure spruce stand on these sites. Our findings suggest that the current FTG management approach in northeast BC only has a positive result if taken to an extreme. It has a low return on investment and reduces stand resilience and total productivity. An alternative forest management approach for the region is presented. Full article