Search Results (1,058)

We propose the HOSA algorithm to pick P-wave arrival times on seismic arrays. HOSA comprises two stages: a single-trace stage (STS) and a multi-channel stage (MCS). STS seeks deviations in higher-order statistics from background noise to identify sets of potential onsets on each trace. STS employs various thresholds and identifies an onset only for solutions that are gently variable with the threshold. Uncertainty is assigned to onsets based on their variation with the threshold. MCS verifies that detected onsets are consistent with the array geometry. It groups onsets within an array by hierarchical agglomerative clustering and selects only groups whose maximum differential times are consistent with the P-wave travel time across the array. HOSA needs a set of P-onsets to be calibrated. These sets may be already available (e.g., preliminary catalogs) or retrieved from picking (manually/automatically) a subset of traces in the target area. We tested HOSA on 226 microearthquakes recorded by 20 temporary arrays of 10 stations each, deployed in the Irpinia region (Southern Italy), which, in 1980, experienced a devastating 6.9 Ms earthquake. HOSA parameters were calibrated using a preliminary catalog of onsets obtained using an automatic template-matching approach. HOSA solutions are more reliable, less prone to false detection, and show higher inter-array consistency than template-matching solutions. Full article

Modeling distributed parameter systems (DPSs) is challenging due to their strong nonlinearity and spatiotemporal coupling. In this study, a three-dimensional fuzzy modeling method combining genetic algorithm (GA)-based automatic clustering and hierarchical extreme learning machine (HELM) is proposed for DPS modeling. The method utilizes GA-based automatic clustering to learn the premise part of 3D fuzzy rules, while HELM is employed to learn spatial basis functions and construct a complete fuzzy rule base. This approach effectively captures the spatiotemporal coupling characteristics of the system and mitigates the information loss commonly observed in dimensionality reduction in traditional fuzzy modeling methods. Through experimental verification, the proposed method is successfully applied to a rapid thermal chemical vapor deposition system. The experimental results demonstrate that the method can accurately predict temperature distribution and maintain good robustness under noise and disturbances. Full article

Transformer models produce advanced text representations that have been used to break through the hard challenge of natural language understanding. Using the Transformer’s attention mechanism, which acts as a language learning memory, trained on tens of billions of words, a word sense disambiguation (WSD) algorithm can now construct a more faithful vectorial representation of the context of a word to be disambiguated. Working with a set of 34 lemmas of nouns, verbs, adjectives and adverbs selected from the National Reference Corpus of Romanian (CoRoLa), we show that using BERT’s attention heads at all hidden layers, we can devise contextual vectors of the target lemma that produce better clusters of lemma’s senses than the ones obtained with standard BERT embeddings. If we automatically translate the Romanian example sentences of the target lemma into English, we show that we can reliably infer the number of senses with which the target lemma appears in the CoRoLa. We also describe an unsupervised WSD algorithm that, using a Romanian BERT model and a few example sentences of the target lemma’s senses, can label the Romanian induced sense clusters with the appropriate sense labels, with an average accuracy of 64%. Full article

Motion primitives are a highly useful and widely employed tool in the field of Learning from Demonstration (LfD). However, obtaining a large number of motion primitives can be a tedious process, as they typically need to be generated individually for each task to be learned. To address this challenge, this work presents an algorithm for acquiring robotic skills through automatic and unsupervised segmentation. The algorithm divides tasks into simpler subtasks and generates motion primitive libraries that group common subtasks for use in subsequent learning processes. Our algorithm is based on an initial segmentation step using a heuristic method, followed by probabilistic clustering with Gaussian Mixture Models. Once the segments are obtained, they are grouped using Gaussian Optimal Transport on the Gaussian Processes (GPs) of each segment group, comparing their similarities through the energy cost of transforming one GP into another. This process requires no prior knowledge, it is entirely autonomous, and supports multimodal information. The algorithm enables generating trajectories suitable for robotic tasks, establishing simple primitives that encapsulate the structure of the movements to be performed. Its effectiveness has been validated in manipulation tasks with a real robot, as well as through comparisons with state-of-the-art algorithms. Full article

The dry bulk shipping network is an important carrier of global bulk commodity flow. To better understand the structural characteristics and future development trends of the global dry bulk shipping network (GDBSN), this study proposes a framework for characteristics analysis and link prediction based on complex network theory. The study integrates large-scale heterogeneous data, including automatic identification system data and port geographic information, to construct the GDBSN. The findings reveal that the network exhibits small-world properties, with the Port of Singapore identified as the most influential node. Link prediction results indicate that many potential new shipping routes exist within regions or between neighboring countries, exhibiting clear regional clustering characteristics. The added links mainly influence the local structure, with minimal impact on the overall network topology. This study provides valuable insights for shipping companies in route planning and for port authorities in developing strategic plans. Full article

Background/Objectives: Level 1 autism spectrum disorder (ASD) is a neurodevelopmental condition characterised by challenges in social and communication skills. Despite these difficulties, individuals with level 1 ASD often exhibit average intelligence and typical language development. Improving socialisation skills in this population requires tailored approaches that address their specific needs and include targeted strategies. This study aims to evaluate the effectiveness of a structured social skills training programme for adolescents and young adults with level 1 ASD. Methods: Participants diagnosed with level 1 ASD, regardless of gender, were consecutively recruited from an outpatient clinic. The intervention involved activities from the Social Skills, Autonomy, and Awareness Module, specifically designed for adolescents and young adults. Sessions were conducted fortnightly, lasting 1.5 to 3 h each, over 17 months. Adaptive behaviour was assessed using the Vineland Adaptive Behaviour Scales (VABS) at baseline and after completing the programme. Data were analysed with SPSS version 22.0 (SPSS Inc., Chicago, IL, USA). Statistical methods included automatic clustering to identify optimal clusters and Pearson’s Chi-square and Fisher’s exact tests to evaluate variable distributions among the clusters. Results: A total of 31 participants (77.4% female) with a mean age of 20.1 years (SD = 7.0) were included in the study. Two distinct clusters emerged. Cluster 1 (n = 8) had significantly higher mean ages and baseline Vineland II socialisation scores than Cluster 2 (n = 23). Both clusters demonstrated significant improvements in social skills following the intervention. Conclusions: This study highlights distinct profiles within individuals with level 1 ASD, showing a clear link between age and social skill development. The intervention improved social skills for most participants, regardless of the age at which treatment began. For some individuals, alternative or augmented treatment strategies may be necessary to achieve optimal results. Full article

Precision agriculture has recently experienced significant advancements through the use of technologies such as unmanned aerial vehicles (UAVs) and satellite imagery, enabling more efficient and precise agricultural management. Yield estimation from these technologies is essential for optimizing resource allocation, improving harvest logistics, and supporting decision-making for sustainable vineyard management. This study aimed to evaluate grape cluster numbers estimated by using YOLOv7x in combination with images obtained by UAVs from a vineyard. Additionally, the capability of several vegetation indices calculated from Sentinel-2 and PlanetScope satellites to estimate grape clusters was evaluated. The results showed that the application of the YOLOv7x model to RGB images acquired from UAVs was able to accurately predict grape cluster numbers (R² value and RMSE value of 0.64 and 0.78 clusters vine⁻¹). On the contrary, vegetation indexes derived from Sentinel-2 and PlanetScope satellites were found not able to predict grape cluster numbers (R² lower than 0.23), probably due to the fact that these indexes are more related to vegetation vigor, which is not always related to yield parameters (e.g., cluster number). This study suggests that the combination of high-resolution UAV images, multispectral satellite images, and advanced detection models like YOLOv7x can significantly improve the accuracy of vineyard management, resulting in more efficient and sustainable agriculture. Full article

Automatic Identification System (AIS) data offer essential insights into maritime traffic patterns; however, effective visualization tools for decision-making remain limited. This study presents an integrated visualization processing method to support ship operators by identifying maritime traffic behavior information, such as traffic density, direction, and flow in specific sea navigational areas. We analyzed AIS dynamic data from a specific sea area, calculated ship density distributions across a grid lattice, and obtained visualizations of traffic-dense areas as heat maps. Using the density-based spatial clustering of applications with a noise algorithm, we detected traffic direction at each grid point, which was visualized in the form of directional arrows, and clustered ship trajectories to identify representative traffic flows. The visualizations were integrated and overlaid onto an S-57-based electronic nautical map for Mokpo’s entry and exit routes, revealing primary shipping lanes and critical inflection points within the target area. This integrated visualization method simultaneously displays traffic density, flow, and customary routes. It is adapted for the electronic nautical chart (S-101) under the next-generation hydrographic information standard (S-100), which can be used as a tool to support decision-making for ship operators. Full article

Currently, the world is rapidly advancing in terms of the construction of new power systems, and planning suitable distribution network planning while also considering renewable energy has become a hot issue. Based on this background, this paper studies the distribution network planning problem. Compared with the traditional planning method, the paper considers the impact of load growth and renewable energy penetration and uses the multi-stage planning method to build the planning model; at the same time, in the scenarios selection, the affinity propagation (AP) clustering algorithm is adopted, which can automatically obtain the number of clusters. Based on the proposed model, an IEEE 33-node is used for simulation. The simulation results show that, compared with the traditional static planning method, the total economic cost of the proposed method is reduced by 4.87% and the wind–solar curtailment rate is reduced by 59.01%; in addition, according to the proposed method, the impact of energy storage equipment and wind–solar permeability on the planning results is studied. It is found that, when considering energy storage, the amount of abandoned wind and light decreases by 22.35% and the total cost first decreases and then increases with the increase in wind–solar permeability, while the total economic cost reaches the minimum at about 40%. The impact of load growth rate on the planning results is also studied. Finally, the generalizability of the proposed method is investigated while using the IEEE 69-node system as an example. Full article

Background: Few studies have evaluated long-COVID in adolescents. Methods: Cohort study. Demographics, clinical data, and the presence of 30 symptoms were collected with a modified WHO form. Mean values were compared by Student’s t test and proportions by the chi-square test or Fisher test, with trends over time analysed using the chi-square test for trend. Potential risk factors independently associated with persisting symptoms were evaluated in a multivariable logistic regression model. Clustering of cases was analysed by two-step automatic clustering. Results: A total of 97 adolescents aged 12–17 (54.6% females, 45.4% males) were evaluated. After a mean interval of 96 days (SD 52) from acute infection, the mean number of symptoms (2.8 overall) was higher for pre-Omicron (3.2 vs. 2.5 in Omicron, p = 0.046) and moderate/severe acute infections (4.2 vs. 2.7 in mild, p = 0.023). Fatigue (62.9%) and dyspnea (43.3%) were the most common symptoms, followed by headache (28.9%), thoracic pain (22.7%), diarrhea (20.6%), palpitations/tachycardia (17.5%), articular pain (15.5%), difficult concentration (14.4%), muscle pain (12.4%), taste reduction (8.2%), smell reduction (8.2%), fever (6.2%), and skin disorders (5.2%). The symptom profile was similar in males and females but showed significant differences from that observed in concurrently followed adults. After a mean interval of 340 days from infection, 45.3% still presented symptoms, with persistence associated with higher number of initial symptoms. Two clusters were defined that differed in the phase of acute infection and the number and profile of symptoms. Conclusions: Long-COVID manifestations in adolescents may differ from those observed in adults. Polisymptomaticity may predict long-term persistence. Full article

Intelligent ships are a key focus for the future development of maritime transportation, relying on efficient decision-making and autonomous control within complex environments. To enhance the perception, prediction, and decision-making capabilities of these ships, the present study proposes a novel approach for constructing a time-series knowledge graph, utilizing real-time Automatic Identification System (AIS) data analyzed via a sliding window technique. By integrating advanced technologies such as knowledge extraction, representation learning, and semantic fusion, both static and dynamic navigational data are systematically unified within the knowledge graph. The study specifically targets the extraction and modeling of critical events, including variations in ship speed, course changes, vessel encounters, and port entries and exits. To evaluate the urgency of encounters, mathematical algorithms are applied to the Distance to Closest Point of Approach (DCPA) and Time to Closest Point of Approach (TCPA) metrics. Furthermore, the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) clustering algorithm is employed to identify suitable docking berths. Additionally, multi-source meteorological data are integrated with ship dynamic data, providing a more comprehensive representation of the maritime environment. The resulting knowledge system effectively combines ship attributes, navigational status, event relationships, and environmental factors, thereby offering a robust framework for supporting intelligent ship operations. Full article

Addressing the limitations of current multi-scale seabed terrain construction methods for a Digital Depth Model (DDM) and the low computational efficiency of automatic generalization algorithms, this paper draws on the concept of curvature simplification from 3D point cloud data processing and proposes a block-based DDM automatic generalization method that leverages surface curvature features. Initially, a clustering blocking model is established using an improved K-means algorithm for partitioning DDM data. Subsequently, a fitting surface is constructed based on the neighboring depth points within the blocked DDM to obtain the surface curvature characteristics of each depth point, which serve as the criterion for the DDM automatic generalization process. By integrating a multi-threaded parallel computation model, an efficient automated generalization workflow that encompasses data partitioning, fitting, computation, processing, and integration of the DDM is ultimately constructed. Furthermore, this paper designs validity and comparative experiments to analyze the proposed algorithm through experimental analysis. The results demonstrate that the algorithm can be applied to the multi-scale construction of DDM seabed terrain, while maintaining the integrity of both flat and complex seabed landforms, and significantly enhancing the computational efficiency of the DDM automatic generalization process. Full article

Deep convolutional networks often encounter information bottlenecks when extracting land object features, resulting in critical geometric information loss, which impedes semantic segmentation capabilities in complex geospatial backgrounds. We developed LULC-SegNet, a semantic segmentation network for land use and land cover (LULC), which integrates features from the denoising diffusion probabilistic model (DDPM). This network enhances the clarity of the edge segmentation, detail resolution, and the visualization and accuracy of the contours by delving into the spatial details of the remote sensing images. The LULC-SegNet incorporates DDPM decoder features into the LULC segmentation task, utilizing machine learning clustering algorithms and spatial attention to extract continuous DDPM semantic features. The network addresses the potential loss of spatial details during feature extraction in convolutional neural network (CNN), and the integration of the DDPM features with the CNN feature extraction network improves the accuracy of the segmentation boundaries of the geographical features. Ablation and comparison experiments conducted on the Circum-Tarim Basin Region LULC Dataset demonstrate that the LULC-SegNet improved the LULC semantic segmentation. The LULC-SegNet excels in multiple key performance indicators compared to existing advanced semantic segmentation methods. Specifically, the network achieved remarkable scores of 80.25% in the mean intersection over union (MIOU) and 93.92% in the F1 score, surpassing current technologies. The LULC-SegNet demonstrated an IOU score of 73.67%, particularly in segmenting the small-sample river class. Our method adapts to the complex geophysical characteristics of remote sensing datasets, enhancing the performance of automatic semantic segmentation tasks for land use and land cover changes and making critical advancements. Full article

The pivotal role of sleep has led to extensive research endeavors aimed at automatic sleep stage classification. However, existing methods perform poorly when classifying small groups or individuals, and these results are often considered outliers in terms of overall performance. These outliers may introduce bias during model training, adversely affecting feature selection and diminishing model performance. To address the above issues, this paper proposes an ensemble-based sequential convolutional neural network (E-SCNN) that incorporates a clustering module and neural networks. E-SCNN effectively ensembles machine learning and deep learning techniques to minimize outliers, thereby enhancing model robustness at the individual level. Specifically, the clustering module categorizes individuals based on similarities in feature distribution and assigns personalized weights accordingly. Subsequently, by combining these tailored weights with the robust feature extraction capabilities of convolutional neural networks, the model generates more accurate sleep stage classifications. The proposed model was verified on two public datasets, and experimental results demonstrate that the proposed method obtains overall accuracies of 84.8% on the Sleep-EDF Expanded dataset and 85.5% on the MASS dataset. E-SCNN can alleviate the outlier problem, which is important for improving sleep quality monitoring for individuals. Full article

The color of a cityscape plays a significant role in its atmosphere; however, the traditional city color analysis methods cover a wide range but are not precise enough, requiring field sampling, a lot of manual comparisons, and lacking quantitative analysis of color. With the development of artificial intelligence, deep learning and computer vision technology show great potential in urban environment research. In this document, we focus on “building color” and present a deep learning-based framework that combines geospatial big data with AI technology to extract and analyze urban color information. The framework is composed of two phases: “deep learning” and “quantitative analysis.” In the “deep learning” phase, a deep convolutional neural network (DCNN)-based color extraction model is designed to automatically learn building color information from street view images; in the “quantitative analysis” phase, building color is quantitatively analyzed at the overall and local levels, and a color clustering model is designed to quantitatively display the color relationship to comprehensively understand the current status of urban building color. The research method and results of this paper are one of the effective ways to combine geospatial big data with GeoAI, which is helpful to the collection and analysis of urban color and provides direction for the construction of urban color information management. Full article