Type Classification Method Research Articles

Interpretation of the Jiangnan Landscape and Countryside (Shan-Shui) Pattern: Evidence from the Classification and Spatial Form of Traditional Settlements in the Nanxi River Basin

Against the backdrop of accelerated globalization and urbanization, traditional settlements in the Jiangnan waterside areas of China face challenges such as morphological distortion and a simplified spatial structure. The ecological adaptability and cultural value of settlements urgently need scientific protection. There has not been enough research on how to better classify settlements, how to conduct systematic morphological analysis, or how to use dynamic protection methods. This makes it hard to fully show the variety of settlement types and how they differ in different areas. To this end, this study uses a combination of GIS spatial analysis, type classification, and case study methods to classify and morphologically analyze 159 traditional settlement patches in the Nanxi River Basin. We construct a settlement type map from quantitative research on topographical, water system, and spatial morphological characteristics, and extract a spatial organization model of mountains and water. The results show the following: (1) The main morphological types, such as the plain–waterfront–strip-shaped and cluster-shaped/finger-shaped types, are typical patterns of traditional settlements in the basin. This indicates how well settlements can adapt to their natural surroundings. (2) This study summarizes six typical settlement sample spaces. (3) The settlement digital protection strategy suggested in this paper uses GIS and 3D modeling technology to make it easier to record, show, and manage information about settlement spaces. This offers a new way to protect traditional settlements scientifically. The study not only enriches the theoretical understanding of the settlement morphology of Jiangnan landscape pastoral areas but also provides an important reference for the protection and sustainable development of settlements in similar basins around the world.

Seismic Risk Classification of Building Clusters Using MST Clustering and UAV Remote Sensing

The fundamental attribute that is essential for the seismic capacity assessment of houses is the building structure type. Conventionally, remote sensing assessment of the seismic capacity for houses has been based on the image features of individual houses, instead of the spatial similarity between them. To enhance the classification accuracy of house structure types, this work proposes a minimum spanning tree (MST) house clustering structure type classification method based on the spatial similarity of houses. First, the method employs the geometric characteristics of residential buildings to calculate the Gestalt factor that characterizes the visual distance. Subsequently, a Delaunay triangular mesh is constructed to create a proximity map between the houses, with the MST generated using visual distance as the weighting factor. Then, the spatial proximity similarity of house clusters is obtained through pruning. Finally, a support vector machine is employed to categorize the architectural structure of the housing complex, viz., simple houses, brick–concrete houses, and frame houses. This classification is based on the geometric, textural, height, and spatial distribution characteristics of the houses. We have conducted a remote sensing classification experiment of house structure types with Zhushan County, Hubei Province as the study area. The results show that the MST clustering method improves the classification accuracy of brick–concrete houses to 95.4% and the classification accuracy of simple houses to 93.4%. Compared to the single-family-based classification method of building structure types, the classification accuracy of frame-structure buildings is improved to 87%. The Kappa coefficient increased to 0.89. This study significantly improves the classification accuracy of building structure types by introducing spatial similarity. Furthermore, it shows the potential for spatial similarity in classifying building structure types.

Bulk Low-Inertia Power Systems Adaptive Fault Type Classification Method Based on Machine Learning and Phasor Measurement Units Data

Bulk Low-Inertia Power Systems Adaptive Fault Type Classification Method Based on Machine Learning and Phasor Measurement Units Data

Application of bioinformatic tools in cell type classification for single-cell RNA-seq data.

The advancements in single-cell RNA sequencing (scRNAseq) technology have significantly transformed genomics research, enabling the handling of thousands of cells in each experiment. As of now, 32,068 research studies have been cataloged in the Pubmed database. The primary aim of scRNAseq investigations is to identify cell types, understand the antitumor immune response, and identify new and uncommon cell types. Traditional techniques for identifying cell types include microscopy, histology, and pathological characteristics. However, the complexity of instruments and the need for precise experimental design make it difficult to fully capture the overall heterogeneity. Unsupervised clustering and supervised classification methods have been used to solve this task. Supervised cell type classification methods have gained popularity as large-scale, high-quality, well-annotated and more robust results compared to clustering methods. A recent study showed that support vector machine (SVM) gives a high-quality classification performance in different scenarios. In this article, we compare and evaluate the performance of four different kernels (sigmoid, linear, radial, polynomial) of SVM. The results of the experiments on three standard scRNA-seq datasets indicate that SVM with linear and SVM with sigmoid kernel classify the cells more accurately (approx. 99 %) where SVM linear kernel method has remarkably fast computation time and we also evaluate the results using some single cell specific evaluation matrices F-1 score, MCC, AUC value. Additionally, it sheds light on the potential use of kernels of SVM to give underlying information of single-cell RNA-Seq data more effectively.

Fundamental investigation of hair type classification method using deep learning for human hair screening at crime scenes

Fundamental investigation of hair type classification method using deep learning for human hair screening at crime scenes

Non-Destructive Assessment of Stone Heritage Weathering Types Based on Machine Learning Method Using Hyperspectral Data

Abstract. Stone cultural heritage is exposed to various environments, resulting in a diverse range of weathering types. The identification of these weathering types is vital for targeted conservation efforts. In this paper, a weathering type classification method based on hyperspectral imaging technology is proposed. Firstly, fresh sandstones are collected from Yungang Grottoes to carry out the simulated weathering experiments, including freeze-thaw cycles and wet-dry cycles with acid, alkali and salt solutions. Subsequently, the hyperspectral imaging system was used to collect the visible-near-infrared (VNIR) and short-wave infrared (SWIR) images of the sandstone samples with different weathering types and degrees. The surface spectral reflectance of sandstone samples with different weathering types were used as training data, with weathering types serving as the labels. Support vector machine (SVM), K-nearest neighbour (KNN), linear discriminant analysis (LDA) and random forest (RF) were used to establish weathering type classification models. The results show that the SVM model and LDA model based on both VNIR and SWIR spectra exhibit outstanding performance, with a best accuracy of 0.994. The framework proposed in this paper facilitates rapid and non-contact assessment of the weathering types of the superficial layers of stone cultural heritage, thereby supporting more targeted conservation work.

A Parallel-Cascaded Ensemble of Machine Learning Models for Crop Type Classification in Google Earth Engine Using Multi-Temporal Sentinel-1/2 and Landsat-8/9 Remote Sensing Data

The accurate mapping of crop types is crucial for ensuring food security. Remote Sensing (RS) satellite data have emerged as a promising tool in this field, offering broad spatial coverage and high temporal frequency. However, there is still a growing need for accurate crop type classification methods using RS data due to the high intra- and inter-class variability of crops. In this vein, the current study proposed a novel Parallel-Cascaded ensemble structure (Pa-PCA-Ca) with seven target classes in Google Earth Engine (GEE). The Pa section consisted of five parallel branches, each generating Probability Maps (PMs) for different target classes using multi-temporal Sentinel-1/2 and Landsat-8/9 satellite images, along with Machine Learning (ML) models. The PMs exhibited high correlation within each target class, necessitating the use of the most relevant information to reduce the input dimensionality in the Ca part. Thereby, Principal Component Analysis (PCA) was employed to extract the top uncorrelated components. These components were then utilized in the Ca structure, and the final classification was performed using another ML model referred to as the Meta-model. The Pa-PCA-Ca model was evaluated using in-situ data collected from extensive field surveys in the northwest part of Iran. The results demonstrated the superior performance of the proposed structure, achieving an Overall Accuracy (OA) of 96.25% and a Kappa coefficient of 0.955. The incorporation of PCA led to an OA improvement of over 6%. Furthermore, the proposed model significantly outperformed conventional classification approaches, which simply stack RS data sources and feed them to a single ML model, resulting in a 10% increase in OA.

An FCM based weather type classification method considering photovoltaic output and meteorological characteristics and its application in power interval forecasting

AbstractWith the increasing installation of photovoltaic (PV) systems, the impact of their randomness and volatility on power system has become a significant concern. To effectively quantify the uncertainty of PV output, it is crucial to develop reliable PV power interval forecasting technology. However, the complex relationship between PV output and meteorological factors makes it challenging for a single forecasting model to meet the forecasting demand. To solve this problem, this paper proposes a weather classification method that takes into account both PV output and meteorological characteristics. Initially, the relationship between PV output and meteorological factors is analyzed, and weather types are classified using fuzzy c‐means algorithm (FCM). Then, an extreme learning machine (ELM) is employed to establish point forecasting model. By combining kernel density estimation, a PV power generation interval forecasting model is derived. The results demonstrate that the FCM‐ELM model achieves higher forecasting accuracy and narrower interval width compared to traditional ELM models, with accuracy improvement of more than 2%. Additionally, the proposed method outperforms seasonal models with an accuracy improvement of more than 1%. The contribution of this paper includes identifying the limitations of traditional weather classification methods, proposing a novel multi‐model approach for PV interval forecasting.

Lyman-α at Cosmic Noon II: The relationship between kinematics and Lyman-α in z ∼ 2–3 Lyman break galaxies

Abstract We report for the first time a relationship between galaxy kinematics and net Lyman- $\alpha$ equivalent width (net Ly $\alpha$ EW) in star-forming galaxies during the epoch of peak cosmic star formation. Building on the previously reported broadband imaging segregation of Ly $\alpha$ -emitting and Ly $\alpha$ -absorbing Lyman break galaxies (LBGs) at $z\sim2$ (Paper I in this series) and previously at $z\sim3$ , we use the Ly $\alpha$ spectral type classification method to study the relationship between net Ly $\alpha$ EW and nebular emission-line kinematics in samples of $z\sim2$ and $z\sim3$ LBGs drawn from the literature for which matching rest-frame UV photometry, consistently measured net Ly $\alpha$ EWs, and kinematic classifications from integral field unit spectroscopy are available. We show that $z\sim2$ and $z\sim3$ LBGs segregate in colour-magnitude space according to their kinematic properties and Lyman- $\alpha$ spectral type and conclude that LBGs with Ly $\alpha$ dominant in absorption (aLBGs) are almost exclusively rotation-dominated (presumably disc-like) systems, and LBGs with Ly $\alpha$ dominant in emission (eLBGs) characteristically have dispersion-dominated kinematics. We quantify the relationship between the strength of rotational dynamic support (as measured using ${v}_{\mathrm{obs}}/2{\sigma }_{\mathrm{int}}$ and ${v}_{\mathrm{rot}}/{\sigma}_{\mathrm{0}}$ ) and net Ly $\alpha$ EW for subsets of our kinematic sample where these data are available, and demonstrate the consistency of our result with other properties that scale with net Ly $\alpha$ EW and kinematics. Based on these findings, we suggest a method by which large samples of rotation- and dispersion-dominated galaxies might be selected using broadband imaging in as few as three filters and/or net Ly $\alpha$ EW alone. If confirmed with larger samples, application of this method will enable an understanding of galaxy kinematic behaviour over large scales in datasets from current and future large-area and all-sky photometric surveys that will select hundreds of millions of LBGs in redshift ranges from $z\sim2-6$ across many hundreds to thousands of Mpc. Finally, we speculate that the combination of our result linking net Ly $\alpha$ EW and nebular emission-line kinematics with the known large-scale clustering behaviour of Ly $\alpha$ -absorbing and Ly $\alpha$ -emitting LBGs is evocative of an emergent bimodality of early galaxies that is consistent with a nascent morphology-density relation at $z\sim2-3$ .

Evolution of Nonparametric Morphology of Galaxies in the JWST CEERS Field at z ≃ 0.8–3.0

Galaxy morphology is one of the most fundamental ways to describe galaxy properties, but the morphology we observe may be affected by wavelength and spatial resolution, which may introduce systematic bias when comparing galaxies at different redshift. Taking advantage of the broad wavelength coverage from optical to near-IR and the high-resolution NIRCam instrument of the JWST, we measure the nonparametric morphological parameters of a total of 1376 galaxies at z ≃ 0.8–3.0 in the CEERS field through an optimized code called statmorph_csst. We divide our sample into three redshift intervals and investigate the wavelength and redshift dependence of the morphological parameters. We also explore how the widely used galaxy type classification methods based on the morphological parameters depend on wavelength and spatial resolution. We find that there are variations in all morphological parameters with rest-frame wavelength (λ rf), especially at the short-wavelength end, and that λ rf mainly affects the classification between late- and early-type galaxies. As λ rf increases, the galaxies on the G–M 20 diagram move to the upper left with a slope of −0.23 ± 0.03 on average. We find that spatial resolution mainly affects the merger identification. The merger fraction in F200W resolution can be ≳2 times larger than that in F444W resolution. Furthermore, we compare the morphological parameter evolution of galaxies with different stellar masses. We find that there are differences in the morphological evolution of high- and low-mass (log M * ≥ 10 and 9 < log M * < 10) galaxies in the studied redshift range, which may be caused by their different evolutionary paths.

3D human body modeling with orthogonal human mask image based on multi-channel Swin transformer architecture

The reconstruction based on RGB images of dressed human body lacks the shape information of the human body under clothing, while the naked 3D human body scanning will violate the user's privacy. To overcome these limitations, a new method, based on Swin transformer (Swin-T), for reconstructing 3D human body shape from human orthogonal mask image is proposed. Its core is to express the reconstruction problem as solving regression mapping function. A fast body shape type classification method based on the human front mask is proposed. The regression function is innovatively represented as a piecewise function, with the body shape of the human body as the segmentation criterion. A multi-channel Swin-T architecture is designed, which can not only extract features from front and side mask images, but also their mixed features to construct the regression mapping function. Different body types for different genders are predicted with separate regression function to help estimate an accurate human model. Extensive experimental results show that the proposed method effectively achieves visually realistic and accurate body reconstruction, and significantly outperforms the current state-of-the-art methods. In addition, the classification of body types can compensate for the errors caused by partial clothing laxity in practical applications, which is beneficial for users to obtain a more accurate 3D human model.

Predicting user types with symbolic images: An empirical validation based on two card-sorting studies

In gamification, personalization is considered an important field of research since it can improve user engagement and motivation. Particularly, effective gamified learning needs to strike a balance between gameful design that promotes engagement and does not negatively affect content comprehension and absorption. The existing traditional user-type questionnaires approaches are time-consuming and intrusive, hampering the user’s focus. Previous research in gamification demonstrated how to automate player profile creation through log file analysis of gameful systems instead of traditional approaches, however, psychological archetypes and motivational groups can simplify the process of collecting data to achieve a balanced level of personalization for gamified systems, which are grounded on the user’s profile and preferences which can help with their immersion. To attend to the needs of a wide range of fields that introduce gamification to systems that are not classified as games, a new ludic approach was created to define user types based on user interaction using symbolic images. In this exploratory research, we tested this approach in two card-sorting studies (N=35 and N=19). Our results show that the images can be used to predict the users’ archetypes and motivational groups accurately. Thus, we contribute with validating an image-based user type classification method that is easier to deploy in systems that do not aim to behave as a game but are looking to reap the benefits of personalizing a user’s content and gameful experience where relevant. Our second contribution is delivering guidelines to replicate this validation method to other existing approaches. We expect that with this guideline, we facilitate personalization with other user typologies without disrupting the gameful experience.

Development of Scale Factors for Skin Type Classification

There are no global standards for the classification of skin types and the consistency of classification results is low. It is necessary to develop a consistent skin type classification method that takes into account innate skin characteristics. First, the concepts of important related terms were established using Delphi techniques from which skin type classification factors were derived. Second, we derived the relative importance and complex weight between each factor using AHP techniques. The research results found that priorities for the classification of skin types were in the order of ‘Oil and moisture levels’, ‘Sensitivity level’, ‘Elasticity level’, and ‘pigmentation level’. Meanwhile, among the 33 detailed factors derived from the scale factor of skin type classification, the importance of ‘sebum mass’ was 1st and ‘pore size’ was 2nd. This study is different to previous studies in that the priorities and weights for each factor were quantitatively derived, and skin type classification was centered on innate factors. This study has limitations such as the lack of experts who participated in Delphi and AHP surveys. However, by deriving key measurement factors to classify skin type, beauty industry workers can use systematic criteria to effectively classify skin types for beauty care purposes in a relatively short period of time.

Cell-type classification of cancer single-cell RNA-seq data using the Subsemble ensemble-based machine learning classifier

Background The advent of single-cell RNA sequencing (scRNA-seq) has provided a high-resolution overview of the cellular heterogeneity of different tissue types. Manual cell type annotation of gene expression datasets remains a useful but time-intensive task. Ensemble machine learning methods leverage the predictive power of multiple classifiers and can be applied to classify high-dimensional gene expression data. Here, we present a novel application of the Subsemble supervised ensemble machine learning classifier used to classify novel cells with known cell type labels using gene expression data. Methods First, we tested the classification performance of different pre-processing steps used to normalize and upsample the training dataset for the Subsemble using a colorectal cancer dataset. Second, we conducted a cross-validated performance benchmark of the Subsemble classifier compared to nine other cell type classification methods across five metrics tested, using an acute myeloid leukemia dataset. Third, we conducted a comparative performance benchmark of the Subsemble classifier using a patient-based leave-one-out cross-validation scheme. Rank normalized scores were calculated for each classifier to aggregate performance across multiple metrics. Results The Subsemble classifier performed best when trained on a dataset that was log-transformed then upsampled to generate balanced class distributions. The Subsemble classifier was consistently the top-ranked classifier across five classification performance metrics compared to the nine other baseline classifiers and showed an improvement in performance as the training dataset increased. When tested using the patient-based leave-one-out cross-validation scheme, the Subsemble was the top-ranked classifier based on rank normalized scores. Conclusions Our proof-of-concept study showed that the Subsemble classifier can be used to accurately predict known cell type labels from single-cell gene expression data. The top-ranked classification performance of the Subsemble across two validation datasets, two cross-validation schemes, and five performance metrics motivates future development of accurate ensemble classifiers of scRNA-seq datasets.

An image-based soil type classification method considering the impact of image acquisition distance factor

An image-based soil type classification method considering the impact of image acquisition distance factor

Research on land cover type classification method based on improved MaskFormer for remote sensing images.

High-resolution remote sensing images have the characteristics of wide imaging coverage, rich spectral information and unobstructed by terrain and features. All of them provide convenient conditions for people to study land cover types. However, most existing remote sensing image land cover datasets are only labeled with some remote sensing images of low elevation plain areas, which is highly different from the topography and landscape of highland mountainous areas. In this study, we construct a Qilian County grassland ecological element dataset to provide data support for highland ecological protection. To highlight the characteristics of vegetation, our dataset only includes the RGB spectrum fused with the near-infrared spectrum. We then propose a segmentation network, namely, the Shunted-MaskFormer network, by using a mask-based classification method, a multi-scale, high-efficiency feature extraction module and a data-dependent upsampling method. The extraction of grassland land types from 2 m resolution remote sensing images in Qilian County was completed, and the generalization ability of the model on a small Gaofen Image Dataset (GID) verified. Results: (1) The MIoU of the optimised network model in the Qilian grassland dataset reached 80.75%, which is 2.37% higher compared to the suboptimal results; (2) the optimized network model achieves better segmentation results even for small sample classes in data sets with unbalanced sample distribution; (3) the highest MIOU of 72.3% is achieved in the GID dataset of open remote sensing images containing five categories; (4) the size of the optimized model is only one-third of the sub-optimal model.

Arctic Sea Ice Lead Detection from Chinese HY-2B Radar Altimeter Data

Sea ice thickness is one of the essential characteristics of sea ice. Sea ice lead detection is the key to sea ice thickness estimation from radar altimetry data. This research studies ten different surface type classification methods, including supervised learning, unsupervised learning, and threshold methods, being applied to the HY-2B radar altimeter data collected in October 2019 in the Arctic Ocean. The Sentinel-1 Synthetic Aperture Radar (SAR) images were used for training and validation of the classifiers. Compared with other classifiers, the supervised Bagging ensemble learning classifier showed excellent and robust performance with overall accuracy up to 95.69%. In order to assess the performance of the Bagging classifier in practical applications, lead fractions from January 2019 to March 2021 based on the HY-2B radar altimeter data were mapped using the trained Bagging classifier and compared to the CryoSat-2 L2I data product. The results of the lead fraction showed the monthly variability of ice lead, and the ice lead had a reasonable spatial distribution and was consistent with CryoSat-2 L2I data products. According to these results, the Bagging classifier can provide an essential reference for future studies of Arctic sea ice thickness and sea level estimation from HY-2B radar altimeter data.

A new fault type classification method in the presence of inverter-based resources

Fault type classification is a critical stage in modern distance relays for activating the effective phase and ground distance elements. It is also used for some supplementary functions such as single-pole tripping and reclosing, teleprotection, and fault location. Traditional fault type classifiers are based on the symmetrical or superimposed components of the current signal. However, owing to the unconventional behavior of inverter-based resources (IBRs) compared to the synchronous generators, traditional methods are not reliable in the presence of IBRs. For example, modern grid codes impose some restrictions in terms of injected current during asymmetrical fault which put the accuracy of conventional fault classification methods in danger. Fault resistance could also affect the accuracy of fault type classification method. In this paper, first, the fault current signature of IBRs is analytically derived. Then, impacts of modern grid code and fault resistance on the conventional fault classification are properly evaluated by executing some simulation studies. Finally, a new fault classifier is proposed which is based on symmetrical components of the local voltage and current during asymmetrical faults. The proposed classifier presents a satisfactory degree of security in the faulted phase identification in the presence of IBRs under different host grid codes. Performance of the proposed method is properly analysed with different simulation scenarios such as different fault locations, fault resistances, pre-fault condition, test systems, and host grid codes in PSCAD/EMTDC software.

AREP: an adaptive, machine learning-based algorithm for real-time anomaly detection on network telemetry data

Abnormal behaviour detection is an essential task of real-time monitoring to secure the reliable operation of ICT infrastructures. This paper presents AREP, an adaptive, long short-term memory-based machine learning algorithm for real-time anomaly detection on network telemetry data. AREP is an improved version of Alter-Re^2, the direct predecessor algorithm developed by our research team. AREP introduces automatic tuning of its two key parameters and includes an offset compensation component to increase accuracy. Unfortunately, AREP and its predecessors perform well only on time series showing specific patterns. Thus, we propose also a data type classification method to identify patterns on which AREP performs best. Moreover, we use an extended range of metrics in our performance evaluations, including area under the curve (AUC). AUC computation is based on receiver operating characteristic (ROC) curves. However, generating ROC curves is not straightforward due to the inherent adaptive threshold technique used by AREP and its predecessors, so we had to develop a novel ROC curve generation approach for these algorithms. We show through rigorous experiments that on network time series following specific data patterns AREP overperforms its predecessors and produces similar or even better performance than other state-of-the-art algorithms.

TransCluster: A Cell-Type Identification Method for single-cell RNA-Seq data using deep learning based on transformer.

Recent advances in single-cell RNA sequencing (scRNA-seq) have accelerated the development of techniques to classify thousands of cells through transcriptome profiling. As more and more scRNA-seq data become available, supervised cell type classification methods using externally well-annotated source data become more popular than unsupervised clustering algorithms. However, accurate cellular annotation of single cell transcription data remains a significant challenge. Here, we propose a hybrid network structure called TransCluster, which uses linear discriminant analysis and a modified Transformer to enhance feature learning. It is a cell-type identification tool for single-cell transcriptomic maps. It shows high accuracy and robustness in many cell data sets of different human tissues. It is superior to other known methods in external test data set. To our knowledge, TransCluster is the first attempt to use Transformer for annotating cell types of scRNA-seq, which greatly improves the accuracy of cell-type identification.