Local Dependence Research Articles

SpectroCVT-Net: A convolutional vision transformer architecture and channel attention for classifying Alzheimer’s disease using spectrograms

Dementia arises from various brain-affecting diseases and injuries, with Alzheimer’s disease being the most prevalent, impacting around 55 million people globally. Current clinical diagnosis often relies on biomarkers indicative of Alzheimer’s distinctive features. Electroencephalography (EEG) serves as a cost-effective, user-friendly, and safe biomarker for early Alzheimer’s detection. This study utilizes EEG signals processed with Short-Time Fourier Transform (STFT) to generate spectrograms, facilitating visualization of EEG signal properties. Leveraging the Brainlat database, we propose SpectroCVT-Net, a novel convolutional vision transformer architecture incorporating channel attention mechanisms. SpectroCVT-Net integrates convolutional and attention mechanisms to capture local and global dependencies within spectrograms. Comprising feature extraction and classification stages, the model enhances Alzheimer’s disease classification accuracy compared to transfer learning methods, achieving 92.59 ± 2.3% accuracy across Alzheimer’s, healthy controls, and behavioral variant frontotemporal dementia (bvFTD). This article introduces a new architecture and evaluates its efficacy with unconventional data for Alzheimer’s diagnosis, contributing: SpectroCVT-Net, tailored for EEG spectrogram classification without reliance on transfer learning; a convolutional vision transformer (CVT) module in the classification stage, integrating local feature extraction with attention heads for global context analysis; Grad-CAM analysis for network decision insight, identifying critical layers, frequencies, and electrodes influencing classification; and enhanced interpretability through spectrograms, illuminating key brain wave contributions to Alzheimer’s, frontotemporal dementia, and healthy control classifications, potentially aiding clinical diagnosis and management.

Semantic dependency and local convolution for enhancing naturalness and tone in text-to-speech synthesis

Self-attention-based networks have become increasingly popular due to their exceptional performance in parallel training and global context modeling. However, it may fall short of capturing local dependencies, particularly in datasets with strong local correlations. To address this challenge, we propose a novel method that utilizes semantic dependency to extract linguistic information from the original text. The semantic relationship between nodes serves as prior knowledge to refine the self-attention distribution. Additionally, to better fuse local contextual information, we introduce a one-dimensional convolution neural network to generate the query and value matrices in the self-attention mechanism, taking advantage of the strong correlation between input characters. We apply this variant of the self-attention network to text-to-speech tasks and propose a non-autoregressive neural text-to-speech model. To enhance pronunciation accuracy, we separate tones from phonemes as independent features in model training. Experimental results show that our model yields good performance in speech synthesis. Specifically, the proposed method significantly improves the processing of pause, stress, and intonation in speech.

A Deep Learning Quantile Regression Photovoltaic Power-Forecasting Method under a Priori Knowledge Injection

Accurate and reliable PV power probabilistic-forecasting results can help grid operators and market participants better understand and cope with PV energy volatility and uncertainty and improve the efficiency of energy dispatch and operation, which plays an important role in application scenarios such as power market trading, risk management, and grid scheduling. In this paper, an innovative deep learning quantile regression ultra-short-term PV power-forecasting method is proposed. This method employs a two-branch deep learning architecture to forecast the conditional quantile of PV power; one branch is a QR-based stacked conventional convolutional neural network (QR_CNN), and the other is a QR-based temporal convolutional network (QR_TCN). The stacked CNN is used to focus on learning short-term local dependencies in PV power sequences, and the TCN is used to learn long-term temporal constraints between multi-feature data. These two branches extract different features from input data with different prior knowledge. By jointly training the two branches, the model is able to learn the probability distribution of PV power and obtain discrete conditional quantile forecasts of PV power in the ultra-short term. Then, based on these conditional quantile forecasts, a kernel density estimation method is used to estimate the PV power probability density function. The proposed method innovatively employs two ways of a priori knowledge injection: constructing a differential sequence of historical power as an input feature to provide more information about the ultrashort-term dynamics of the PV power and, at the same time, dividing it, together with all the other features, into two sets of inputs that contain different a priori features according to the demand of the forecasting task; and the dual-branching model architecture is designed to deeply match the data of the two sets of input features to the corresponding branching model computational mechanisms. The two a priori knowledge injection methods provide more effective features for the model and improve the forecasting performance and understandability of the model. The performance of the proposed model in point forecasting, interval forecasting, and probabilistic forecasting is comprehensively evaluated through the case of a real PV plant. The experimental results show that the proposed model performs well on the task of ultra-short-term PV power probabilistic forecasting and outperforms other state-of-the-art deep learning models in the field combined with QR. The proposed method in this paper can provide technical support for application scenarios such as energy scheduling, market trading, and risk management on the ultra-short-term time scale of the power system.

Activities-specific balance confidence scale: a Rasch evaluation of the Arabic version in lower-limb prosthetic users

Purpose To evaluate the psychometric properties of the Arabic version of the Activities-Specific Balance Confidence Scale using the 5-option response categories for individuals with lower limb amputation (ABC-5/Ar). Materials and methods This was a methodological study on a convenience sample of individuals with unilateral lower-limb amputation attending outpatient rehabilitation centres in Saudi Arabia and Turkey (N = 155). Rasch analysis (WINSTEPS version 4.6.5) was used to examine the 5-categories rating scale structure, item fit, item difficulty hierarchy, person separation index, unidimensionality, local item dependency, and differential item functioning. Results The ABC-5/Ar 5-response option demonstrated an appropriate model fit. Most items fit the Rasch model, except for item #12 “walk in a crowded mall,” which showed an overfitting value of 0.63. The person separation indices 2.95 (Cronbach’s α = 0.96). Principal component analysis of residuals confirmed the unidimensionality of the scale; however, local dependency was detected between item #14 “Ride in escalator holding rail,” and item #15 “Ride in escalator not holding rail.” Conclusions The findings suggest that the ABC-5/Ar shows promise in assessing balance confidence in Arabic-speaking lower-limb prosthesis users. However, further studies with larger sample sizes and in diverse clinical contexts are needed to confirm its effectiveness in various clinical settings.

Developing Theoretical Models for Atherosclerotic Lesions: A Methodological Approach Using Interdisciplinary Insights.

Atherosclerosis, a leading cause of cardiovascular disease, necessitates advanced and innovative modeling techniques to better understand and predict plaque dynamics. The present work presents two distinct hypothetical models inspired by different research fields: the logistic map from chaos theory and Markov models from stochastic processes. The logistic map effectively models the nonlinear progression and sudden changes in plaque stability, reflecting the chaotic nature of atherosclerotic events. In contrast, Markov models, including traditional Markov chains, spatial Markov models, and Markov random fields, provide a probabilistic framework to assess plaque stability and transitions. Spatial Markov models, visualized through heatmaps, highlight the spatial distribution of transition probabilities, emphasizing local interactions and dependencies. Markov random fields incorporate complex spatial interactions, inspired by advances in physics and computational biology, but present challenges in parameter estimation and computational complexity. While these hypothetical models offer promising insights, they require rigorous validation with real-world data to confirm their accuracy and applicability. This study underscores the importance of interdisciplinary approaches in developing theoretical models for atherosclerotic plaques.

A multi-scale attributes fusion model for travel mode identification using GPS trajectories

ABSTRACT Travel mode recognition is a key issue in urban planning and transportation research. While traditional travel surveys use manual data collection and have limited coverage, poor timeliness, and insufficient sample capacity, recent advancements in Global Positioning System (GPS) technology allow large-scale data collection and offer novel opportunities to enhance travel mode recognition. However, existing studies often neglect regular differences and changes in motion states across different travel modes and fail to fully integrate multi-scale spatio-temporal features, which limits the accurate classification of travel modes. To fill this gap, this study proposes a multi-scale spatio-temporal attribute fusion (MSAF) model for precise travel mode identification using solely GPS trajectories without altering their sampling rate. The MSAF model segments GPS trajectories into various temporal and spatial scales, extracting local motion states and spatial features at multiple scales. The spatio-temporal feature extraction module is constructed to extract local motion states and capture spatio-temporal dependencies. Additionally, the model incorporates a multi-scale feature fusion module, which effectively combines features of various scales through a series of fusion techniques to obtain a comprehensive representation, enabling automatic and accurate travel mode identification. Experiments on real-world datasets, including the GeoLife Trajectories dataset and the Sussex-Huawei Locomotion-Transportation (SHL) dataset, demonstrate the effectiveness of the MSAF model, achieving a competitive accuracy of 95.16% and 91.70%. This represents an improvement of 2.50% to 7.95% and 0.8% to 6.62% over several state-of-the-art baselines, effectively addressing sample imbalance challenges. Moreover, the experiments demonstrate the significant role of multiscale feature fusion in improving model performance.

Modeling method of local financial dependence: Evidence from Mongolia

Many countries have been striving to equalize the balance between the central government and sub-government financial disparities without considering political interference. This paper aims to summarize the theory of local financial federalism and propose an unprecedented model for identifying the factors that affect local financial dependence. The proposed model is based on a theoretical framework incorporating five assumptions and is applied through panel regression analysis. Utilizing panel data from 2013 to 2022 in 21 provinces of Mongolia, a total of nine variables have been identified and econometrically tested within the proposed model. The findings from the panel regression analysis reveal that local budget investment, local personal income tax, local property tax, and other local taxes positively impact the reduction of local financial dependence. However, it is observed that an increase in local budget expenditures and GDP leads to an escalation in local financial dependence.

SGformer: Boosting transformers for indoor lighting estimation from a single image

Estimating lighting from standard images can effectively circumvent the need for resource-intensive high-dynamic-range (HDR) lighting acquisition. However, this task is often ill-posed and challenging, particularly for indoor scenes, due to the intricacy and ambiguity inherent in various indoor illumination sources. We propose an innovative transformer-based method called SGformer for lighting estimation through modeling spherical Gaussian (SG) distributions—a compact yet expressive lighting representation. Diverging from previous approaches, we explore underlying local and global dependencies in lighting features, which are crucial for reliable lighting estimation. Additionally, we investigate the structural relationships spanning various resolutions of SG distributions, ranging from sparse to dense, aiming to enhance structural consistency and curtail potential stochastic noise stemming from independent SG component regressions. By harnessing the synergy of local-global lighting representation learning and incorporating consistency constraints from various SG resolutions, the proposed method yields more accurate lighting estimates, allowing for more realistic lighting effects in object relighting and composition. Our code and model implementing our work can be found at https://github.com/junhong-jennifer-zhao/SGformer.

GNNctd: A graph neural network based on complicated temporal dependencies modeling for fashion trend prediction

Fashion trend forecasting has consistently remained a focal point within the realm of fashion. Existing methods predominantly concentrate on the external factors influencing fashion trends, often disregarding the intricate interplay among distinct fashion elements, namely the ‘spatial dependencies’ among them. It is also a significant challenge to excavate complicated temporal relationships in complex time series data. In this research, our primary focus is modeling the relationships among diverse fashion elements and the intricate temporal dependencies within time series data. First, we propose a Street Fashion Trend (SFT) dataset by leveraging images from the popular photo-sharing platform Flickr 11Flickr: www.flickr.com. and the renowned fashion show website Vogue.22Vogue: https://m.vogue.com.cn/. Furthermore, we propose a model called GNNctd to solve the above problems. This model leverages a spatial dependency capture module (SDCM) based on a graph neural network to dynamically model the ‘spatial dependency relationships’ among distinct fashion elements. Meanwhile, the model introduces a temporal relationship extraction block (TREB), which comprises two pivotal modules: the interaction learning (IL) module designed to capture local temporal dependencies and a global time attention module (GTAM) that is used to capture global temporal dependencies. Many experiments substantiate that the proposed GNNctd model can achieve more accurate predictions of fashion trends in our constructed dataset and the other three fashion trend datasets. Simultaneously, the GNNctd model achieves a state-of-the-art performance on the Solar-Energy, Exchange Rate, and Wind datasets within the domain of time series prediction.

Efficient multi-level cross-modal fusion and detection network for infrared and visible image

With the rapid development of uncrewed aerial vehicle (UAV) technology, detecting aerial images has found significant applications across various domains. However, existing algorithms overlook the impact of illumination on target detection, resulting in less satisfactory detection performance under low-light conditions. We propose EfficientFuseDet, a visible and infrared image fusion detection network to overcome this issue. First, an effective multilevel cross-modal fusion network called EfficientFuse is presented to combine complementary information from both modalities better. EfficientFuse captures local dependencies and global contextual information in shallow and deep layers, seamlessly combining complimentary local and global features throughout the network. The generated fused images can exhibit clear target contours and abundant texture information. Second, we propose a detection network called AFI-YOLO, which employs an inverted residual vision transformer backbone (IRViT) to effectively address the challenges associated with background interference in fused images. We design an efficient feature pyramid network (EFPN) that efficiently integrates multiscale information, enhancing multiscale detection capability using aerial images. A reparameterization decoupling head (RepHead) is proposed to further improve target classification and localization precision. Finally, experiments on the DroneVehicle dataset indicate that the detection accuracy using fused images can reach 47.2 %, which is higher than that observed with visible light images of 45 %. Compared to state-of-the-art detection algorithms, EfficientFuseDet exhibits a slight decrease in speed. However, it demonstrates superior detection capabilities and effectively enhances the detection accuracy using aerial images under low-light conditions.

Joint extraction of Chinese medical entities and relations based on RoBERTa and single-module global pointer

BackgroundMost Chinese joint entity and relation extraction tasks in medicine involve numerous nested entities, overlapping relations, and other challenging extraction issues. In response to these problems, some traditional methods decompose the joint extraction task into multiple steps or multiple modules, resulting in local dependency in the meantime.MethodsTo alleviate this issue, we propose a joint extraction model of Chinese medical entities and relations based on RoBERTa and single-module global pointer, namely RSGP, which formulates joint extraction as a global pointer linking problem. Considering the uniqueness of Chinese language structure, we introduce the RoBERTa-wwm pre-trained language model at the encoding layer to obtain a better embedding representation. Then, we represent the input sentence as a third-order tensor and score each position in the tensor to prepare for the subsequent process of decoding the triples. In the end, we design a novel single-module global pointer decoding approach to alleviate the generation of redundant information. Specifically, we analyze the decoding process of single character entities individually, improving the time and space performance of RSGP to some extent.ResultsIn order to verify the effectiveness of our model in extracting Chinese medical entities and relations, we carry out the experiments on the public dataset, CMeIE. Experimental results show that RSGP performs significantly better on the joint extraction of Chinese medical entities and relations, and achieves state-of-the-art results compared with baseline models.ConclusionThe proposed RSGP can effectively extract entities and relations from Chinese medical texts and help to realize the structure of Chinese medical texts, so as to provide high-quality data support for the construction of Chinese medical knowledge graphs.

The Flare-OA-16 questionnaire measuring flare in knee and hip osteoarthritis in the patient perspective: scale reduction and validation using a Rasch model

ObjectivesThe recent Flare-OA questionnaire measuring flare in knee and hip osteoarthritis (OA) (19 items in 5 domains, numerical rating scale) showed good psychometric properties along with classical test theory. This study aimed to determine its scaling properties by Rasch analysis and to present evidence for a refined scalable version. Study Design and SettingThe participants were 398 subjects (mean age 64 years [standard deviation = 8.1], 70.4% women) recruited from Australia, France, and the United States, with clinically and radiologically symptomatic knee or hip OA, who completed an online survey. The sample was split into derivation and validation subsamples, stratified by country and joint. Rasch analysis examined differential item functioning (DIF) for sex, age, country, and joint. A confirmatory factor analysis and an analysis of convergent validity were performed to document the psychometric properties of the short version. ResultsTo fit the Rasch model, we reordered thresholds of answering modalities when necessary. Two items were removed. A local dependency between 2 items was solved by combining items modalities into a super-item. A uniform DIF (expected and nonremoved) was identified for one item that was split by joint, and a nonuniform DIF for one item for age and country (removed). The person-item threshold distribution showed a well-focused scale; the confirmatory factor analysis and the analysis of convergent validity showed good fit indicators for the short version. ConclusionThe Rasch analysis was helpful in guiding the decision to refine the measurement instrument. After analysis, the 16-item Flare-OA self-report questionnaire is available for use in clinical research.

A Rotating Machinery Fault Diagnosis Method Based on Dynamic Graph Convolution Network and Hard Threshold Denoising.

With the development of precision sensing instruments and data storage devices, the fusion of multi-sensor data in gearbox fault diagnosis has attracted much attention. However, existing methods have difficulty in capturing the local temporal dependencies of multi-sensor monitoring information, and the inescapable noise severely decreases the accuracy of multi-sensor information fusion diagnosis. To address these issues, this paper proposes a fault diagnosis method based on dynamic graph convolutional neural networks and hard threshold denoising. Firstly, considering that the relationships between monitoring data from different sensors change over time, a dynamic graph structure is adopted to model the temporal dependencies of multi-sensor data, and, further, a graph convolutional neural network is constructed to achieve the interaction and feature extraction of temporal information from multi-sensor data. Secondly, to avoid the influence of noise in practical engineering, a hard threshold denoising strategy is designed, and a learnable hard threshold denoising layer is embedded into the graph neural network. Experimental fault datasets from two typical gearbox fault test benches under environmental noise are used to verify the effectiveness of the proposed method in gearbox fault diagnosis. The experimental results show that the proposed DDGCN method achieves an average diagnostic accuracy of up to 99.7% under different levels of environmental noise, demonstrating good noise resistance.

Examining the Reliability and Validity of the ALS Certification Examinations with the Inclusion of Clinical Judgment: An Update on the ALS Examination Redesign

Objectives Clinical judgment describes the process an emergency medical service clinician uses to evaluate problems and make decisions in the out-of-hospital setting. As part of the redesign of the Advanced Life Support (ALS) certification examinations, the National Registry of Emergency Medical Technicians is developing and evaluating items that measure clinical judgment, with the intention of assessing these as a new domain in the ALS certification examinations. In this study, we provide evidence around the redesign by evaluating the reliability and validity of the advanced emergency medical technician (AEMT) and paramedic certification examinations when clinical judgment is included as a sixth domain along with the five current domains. Methods Pretest (i.e., pilot, unscored) clinical judgment items were included as a new sixth clinical judgment domain. We then used the combination of operational (i.e., scored) and pretest items for all six domains and scored the redesigned AEMT and paramedic certification examinations. We evaluated the psychometric properties of these ALS examinations within the Rasch measurement framework with multiple assessments of reliability and validity including item-level statistics (e.g., mean-square infit and outfit, local dependence) and examination-level statistics (e.g., person reliability, item reliability, item separation, decision consistency, decision accuracy). Wright Maps were produced to evaluate whether the examination item difficulty statistics aligned with the candidate ability continuum. Results The total population of all examination forms included were 20,136 (AEMT 4,983; paramedic 15,153). The Rasch-based statistics for the redesigned AEMT and paramedic examinations, for both item and examination-level statistics, were well within the psychometric standard values. Wright maps demonstrated that the developed items fall along the candidate ability continuum for both examinations. Further, the distribution of clinical judgment item difficulties fell within the current item distribution, providing evidence that these new items are of similar difficulty to the items measuring the five current domains. Conclusion We demonstrate strong reliability and validity evidence to support that the integrity of the examinations is upheld with the addition of clinical judgment items, while also providing a more robust candidate evaluation. Most importantly, the pass/fail decisions that candidates receive accurately reflect their level of ALS knowledge at the entry-level.

Evolution of Displacement Speed Statistics During Head-On Flame-Wall Interaction within Turbulent Boundary Layers

ABSTRACT The statistical behaviours of the density-weighted displacement speed and its curvature and strain rate dependence during head-on interaction of statistically planar premixed flames have been analysed based on Direct Numerical Simulations. The analysis has been conducted within turbulent boundary layers featuring inert walls, under both isothermal and adiabatic wall boundary conditions. The flame quenches due to heat loss when it reaches close to the wall in the case of isothermal wall boundary condition. By contrast, the flame can propagate all the way to the wall before extinguishing due to the consumption of reactants in the case of adiabatic wall boundary conditions. Thus, the effects of thermal expansion, quantified by dilatation rate, and flame normal acceleration during flame-wall interaction in the case of the isothermal wall are weaker than that in the case of the adiabatic wall case due to flame quenching. This gives rise to significant differences in the curvature and strain rate dependences of the reactive scalar gradient magnitude, which affects the statistical behaviours of the reaction and normal diffusion components of density-weighted displacement speed including their mean values, widths of their probability density functions, and their local strain rate and curvature dependences. It has been found that the interaction of near-wall vortical structures with flame surface increases the range of curvature variation during head-on interaction, which acts to widen the probability density function of the density-weighted displacement speed. This trend is relatively stronger for the isothermal wall boundary condition because of the larger variation of the reaction rate component of the density-weighted displacement speed due to local flame quenching, which also acts to widen the range of the density-weighted displacement speed variation in comparison to that in the case of adiabatic wall boundary conditions. Although the qualitative nature of curvature, strain rate, and stretch rate dependences of the density-weighted displacement speed remain unaffected by the wall boundary condition, the strength of the correlation changes with the progress of head-on interaction. It has been found that the negative correlation between the density-weighted displacement speed and flame curvature weakens with the progress of head-on interaction, which gives rise to a reduction in the strength of the correlation between the density-weighted displacement speed and flame stretch rate. Similarly, the correlation between the tangential strain rate and the density-weighted displacement speed weakens with the progress of head-on interaction. Detailed physical explanations are provided for these behaviours, and their modeling implications are indicated.

The chemical ordering and local pressure dependence for truncated octahedron PdAuRh nanoalloys

The optimal chemical ordering configurations of 38-atom trimetallic Pd n Au(32−n)Rh6 nanoalloys with truncated octahedron (TO) geometry were determined through Monte Carlo Basin-Hopping algorithm employing the Gupta potential, followed by local relaxations. Variations in mixing energy were assessed to compare relative stability, revealing the lowest excess energy at Pd16Au16Rh6 composition at the Gupta level. Furthermore, a comprehensive analysis of local atomic pressure was conducted, exploring the factors influencing these pressures.

Linear Logistic Scoring Equations for Latent Class and Latent Profile Models: A Simple Method for Classifying New Cases

Researchers are often interested in using latent class or latent profile parameter estimates to obtain posterior class membership probabilities for observations other than those of the original sample. In this paper, we demonstrate that these probabilities typically take on the form of linear logistic equations with coefficients which are functions of the original model parameters. In other words, the posterior class membership probabilities can be computed with a prediction formula similar to that of a multinomial logistic regression model. We derive the scoring equations for nominal, ordinal, count, and continuous indicators, as well as investigate models with missing values on class indicators, local dependencies, covariates, or multiple latent variables. In addition to the mathematical derivations of the scoring equations, we describe how either exact or approximate scoring equations can be obtained by estimating a multinomial regression model using a weighted data set.

Accuracy of discrete- and continuous-time mean-field theories for epidemic processes on complex networks.

Discrete- and continuous-time approaches are frequently used to model the role of heterogeneity on dynamical interacting agents on the top of complex networks. While, on the one hand, one does not expect drastic differences between these approaches, and the choice is usually based on one's expertise or methodological convenience, on the other hand, a detailed analysis of the differences is necessary to guide the proper choice of one or another approach. We tackle this problem by investigating both discrete- and continuous-time mean-field theories for the susceptible-infected-susceptible (SIS) epidemic model on random networks with power-law degree distributions. We compare the discrete epidemic link equations(ELE) and continuous pair quenched mean-field (PQMF) theories with the corresponding stochastic simulations, both theories that reckon pairwise interactions explicitly. We show that ELE converges to the PQMF theory when the time step goes to zero. We performed an epidemic localization analysis considering the inverse participation ratio (IPR). Both theories present the same localization dependence on network degree exponent γ: for γ<5/2 the epidemics are localized on the maximum k-core of networks with a vanishing IPR in the infinite-size limit while, for γ>5/2, the localization happens on hubs that do not form a densely connected set and leads to a finite value of the IPR. However, the IPR and epidemic threshold of ELE depend on the time-step discretization such that a larger time step leads to more localized epidemics. A remarkable difference between discrete- and continuous-time approaches is revealed in the epidemic prevalence near the epidemic threshold, in which the discrete-time stochastic simulations indicate a mean-field critical exponent θ=1 instead of the value θ=1/(3-γ) obtained rigorously and verified numerically for the continuous-time SIS on the same networks.

Geospatial analysis of scour development post-protection in offshore wind farm

Scour is one of the most serious challenges affecting the normal operation of offshore wind power. Whether scour protection measures are effective directly affects the safety and economical efficiency of an offshore wind farm. Data mining techniques and analytical tools based on monitoring data analysis plays a critical role in making appropriate strategic decisions. Post-protection scour inspection was continuously carried out on a typical offshore wind farm composed of 72 turbines installed with monopile foundations for three times over the course of two years. The geospatial properties of scour, including maximum scour depth, maximum scour extension, and scouring and siltation volume, were analyzed in detail. The possible autocorrelation of scour data measured for each turbine was identified using data-driven decision-making process. The results indicate that the characteristics of the investigated scour and their spatial dependence on the turbines of the inspected offshore wind farm generally show an upward trend over time. A High-High area in the local geospatial dependency map of maximum scour depth was identified, and an equation was drawn for predicting development of maximum scour depth. These findings show indications to reduce the operation and maintenance costs of offshore wind farms and improve the efficiency for upgrading scour protection on specific offshore wind foundations.

TransLSTM: A hybrid LSTM-Transformer model for fine-grained suggestion mining

Digital platforms on the internet are invaluable for collecting user feedback, suggestions, and opinions about various topics, such as company products and services. This data is instrumental in shaping business strategies, enhancing product development, and refining service delivery. Suggestion mining is a key task in natural language processing, which focuses on extracting and analysing suggestions from these digital sources. Initially, suggestion mining utilized manually crafted features, but recent advancements have highlighted the efficacy of deep learning models, which automatically learn features. Models like Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM), and Bidirectional Encoder Representations from Transformers (BERT) have been employed in this field. However, considering the relatively small datasets and the faster training time of LSTM compared to BERT, we introduce TransLSTM, a novel LSTM-Transformer hybrid model for suggestion mining. This model aims to automatically pinpoint and extract suggestions by harnessing both local and global text dependencies. It combines the sequential dependency handling of LSTM with the contextual interaction capabilities of the Transformer, thus effectively identifying and extracting suggestions. We evaluated our method against state-of-the-art approaches using the SemEval Task-9 dataset, a benchmark for suggestion mining. Our model shows promising performance, surpassing existing deep learning methods by 6.76% with an F1 score of 0.834 for SubTask A and 0.881 for SubTask B. Additionally, our paper presents an exhaustive literature review on suggestion mining from digital platforms, covering both traditional and state-of-the-art text classification techniques.