Feature Clustering Research Articles

Assessing Safety Efficiency in China’s Provincial Construction Industry: Trends, Influences, and Implications

Ensuring safety is crucial for promoting the sustainable growth of the construction industry. Assessing safety efficiency is of significant importance for optimizing safety management processes and improving the safety environment. However, the current mainstream methods for evaluating safety efficiency have limitations such as ignoring non-desired outputs and slack variables, the efficiency values being limited to the (0, 1) range, and a narrow perspective. To address these shortcomings, this study focuses on the characteristics of the construction industry and introduces the Super-SBM model and Malmquist index into the assessment of safety efficiency in the construction industry. The study analyzes the evolution characteristics of safety efficiency from both static and dynamic perspectives. Furthermore, using panel quantile regression models, the study identifies the factors influencing safety efficiency and analyzes their heterogeneity. Analyzing panel data from 30 provinces in China from 2015 to 2021, the results show that the overall safety efficiency of the construction industry in China is relatively low, with noticeable spatial clustering characteristics. Provinces in the eastern and central regions exhibit higher levels of construction safety efficiency. The Malmquist index demonstrates a declining trend, with technical efficiency being the primary factor limiting the improvement of safety efficiency in construction. Factors such as per capita GDP, urbanization rate, committed contract amounts, and the number of professionals engaged in survey and design, as well as engineering supervision, have an impact on construction safety efficiency, and the effects of these variables vary across different quantile levels of safety efficiency. This research can assist decision-makers in gaining a better understanding of the safety conditions in different regions of the construction industry. It can also assist in developing customized policies to enhance the health and safety environment, thereby promoting the stable development of the construction industry.

Особенности использования кластерного подхода при организации государственно-частного партнерства в сфере туризма

The study of the features of using the cluster approach in organizing public-private partnerships in the field of tourism is highly relevant in the context of the continued sanctions regime against Russia by the United States and the European Union. This is due, on the one hand, to the multiple growth of domestic tourism in many regions of Russia, and on the other, to high geopolitical risks and limited financial resources of Russian business. The article describes the dynamics of the development of public-private partnerships and its organizational features in the Russian Federation. In particular, the absolute priority of the share of concession agreements in the overall structure of public-private partnership projects in 2022 is shown; the structural differences between the initiatives of public-private partnership projects in Europe and the Russian Federation in 2020-2022 are identified, as well as the national characteristics of public-private partnerships in relation to the tourism complex, the organization of which is still a poorly developed method of improving national tourism. A significant difference between a domestic organization and a foreign one is that in Russia tourism is considered as a separate sector, while in Europe public-private partnership projects indirectly influence the improvement of tourism in the regions through the development of transport and social infrastructures. Whereas in Russia, in addition to projects of indirect importance, much attention is paid to potentially useful objects for increasing direct endogenous tourism indicators, such as the number of tourist arrivals, the number of foreign arrivals, increasing the share of tourism in GDP, passenger traffic, etc. It is concluded that a characteristic feature of the development of the tourism complex in the public-private partnership (PPP) format can be considered the use of the cluster approach as the main one. This allowed the authors to reveal a set of advantages of PPP in tourism; formulate the characteristics of a tourism cluster formed through the association of representatives of government agencies and private business; propose a scheme for the effectiveness of PPP design in tourism using a cluster approach.

Effects of green pruning and combine microelement applications on bud fruitfulness, vegetative development and cluster characteristics of ‘Trakya İlkeren’ grape cultivar

Green pruning practices are frequently used in viticulture. These practices provide better light penetration, successful air circulation, prevents air humidity and aids disease control in the grapevine canopy. However, excessive leaf removal inhibits vine growth and good fruit ripening. In this study, the effects of some foliar and canopy management practices on cluster, vegetative growth traits and bud fruitfulness of the following year were investigated. Some green pruning (unproductive shoot removal, leaf removal, shoot tip removal, shoot topping), foliar microelements and green pruning + microlements treatments were applied in ‘Trakya İlkeren’ grape cultivar. Microelement application increased cluster (382.1 g) and berry weight (4.4 g), berry length (19.7 mm) and width (18.2 mm), cluster length (19.5 cm) and width (13.9 cm), berry volume (3.30 cm3) and berry firmness (7.46 N). The highest maturity index was detected in the control group (23.4). While the most intense L* color value was determined in the green pruning + micro element application, the b* color value was determined in the green pruning application. It was determined that leaf area (211.2 cm2), leaf chlorophyll index (32.75 SPAD), shoot diameter (11.13 mm), shoot length (194.71 cm) and internode length (76.81 mm) increased with microelement application. Among the treatments, the most effective application on bud fruitfulness in the following year was at microelement application (1.34 clusters). The effects of the treatments on the amount of Total soluble solids pH, titratable acid and a* color value were not significant. In this study; green pruning + microelement application was recommended for yield and quality sustainability. According to bud fruitfulness results, the highest cluster numbers per node were obtained from pruning with 4 to 10 buds. It is recommended to prune medium or long to obtain higher yields from this cultivar.

Cross-Domain Contrastive Learning for Time Series Clustering

Most deep learning-based time series clustering models concentrate on data representation in a separate process from clustering. This leads to that clustering loss cannot guide feature extraction. Moreover, most methods solely analyze data from the temporal domain, disregarding the potential within the frequency domain. To address these challenges, we introduce a novel end-to-end Cross-Domain Contrastive learning model for time series Clustering (CDCC). Firstly, it integrates the clustering process and feature extraction using contrastive constraints at both cluster-level and instance-level. Secondly, the data is encoded simultaneously in both temporal and frequency domains, leveraging contrastive learning to enhance within-domain representation. Thirdly, cross-domain constraints are proposed to align the latent representations and category distribution across domains. With the above strategies, CDCC not only achieves end-to-end output but also effectively integrates frequency domains. Extensive experiments and visualization analysis are conducted on 40 time series datasets from UCR, demonstrating the superior performance of the proposed model.

SEC: More Accurate Clustering Algorithm via Structural Entropy

As one of the most popular machine learning tools in the field of unsupervised learning, clustering has been widely used in various practical applications. While numerous methods have been proposed for clustering, a commonly encountered issue is that the existing clustering methods rely heavily on local neighborhood information during the optimization process, which leads to suboptimal performance on real-world datasets. Besides, most existing clustering methods use Euclidean distances or densities to measure the similarity between data points. This could constrain the effectiveness of the algorithms for handling datasets with irregular patterns. Thus, a key challenge is how to effectively capture the global structural information in clustering instances to improve the clustering quality. In this paper, we propose a new clustering algorithm, called SEC. This algorithm uses the global structural information extracted from an encoding tree to guide the clustering optimization process. Based on the relation between data points in the instance, a sparse graph of the clustering instance can be constructed. By leveraging the sparse graph constructed, we propose an iterative encoding tree method, where hierarchical abstractions of the encoding tree are iteratively extracted as new clustering features to obtain better clustering results. To avoid the influence of easily misclustered data points located on the boundaries of the clustering partitions, which we call "fringe points", we propose an iterative pre-deletion and reassignment technique such that the algorithm can delete and reassign the "fringe points" to obtain more resilient and precise clustering results. Empirical experiments on both synthetic and real-world datasets demonstrate that our proposed algorithm outperforms state-of-the-art clustering methods and achieves better clustering performances. On average, the clustering accuracy (ACC) is increased by 1.7% and the normalized mutual information (NMI) by 7.9% compared with the current state-of-the-art (SOTA) algorithm on synthetic datasets. On real-world datasets, our method outperforms other clustering methods with an average increase of 12.3% in ACC and 5.2% in NMI, respectively.

Abstract 2587: Evaluating the potential of radiomics features to predict outcome to an antibody-drug conjugate therapy in non-small cell lung cancer patients

Abstract This study evaluated radiomics, a non-invasive biomarker, for assessment of tumor heterogeneity in non-small cell lung cancer (NSCLC) patients treated with an AXL-targeting antibody-drug conjugate (ADC), enapotamab vedotin (EnaV), through analyzing baseline and post-treatment radiomic features from CT images for predicting outcome. Clinical development of EnaV was discontinued in 2020 due to low response rates unimproved by dose optimization and/or predictive biomarkers. Study data included CT images, target expression and clinical data from 77 advanced NSCLC patients treated at 1.0 mg/kg (3Q4W), 1.8 mg/kg (Q3W), or 2.2 mg/kg EnaV (1Q3W) in the first in human phase I/II study assessing safety and preliminary efficacy of EnaV (NCT02988817). Radiologist-segmented tumors were examined using the cancer imaging phenomics tooklit CaPTk1, deriving 225 radiomic features from 77 baseline and 50 Cycle 2 post-treatment CT images. Batch effects were balanced using nested ComBat2, and radiomic feature dimensionality was reduced in two stages: clustering features with ≥80% similarity followed by principal component analysis (PCA). Radiomic PC1 was utilized in statistical models to assess overall and progression-free survival (OS, PFS), overall response rate (ORR), and disease control rate (DCR). Both baseline and delta radiomic PC1 were evaluated for correlation with biomarker and clinical activity, independently and with soluble or tumoral AXL expression. Baseline radiomic PC1 significantly predicted DCR: confirmed (p=0.03), and unconfirmed (p=0.03). Baseline tumor membrane AXL predicted ORR for this NSCLC subset: confirmed (p=0.03) or unconfirmed (p=0.02). Delta analysis showed delta PC1 predicted unconfirmed (p=0.03) and confirmed (p=0.03) DCR and remained predictive for DCR with soluble AXL (p=0.04). No association was found between baseline and delta radiomics with OS or PFS. Baseline and delta radiomic signatures showed potential in predicting NSCLC response. Preliminary findings suggest that radiomics combined with target expression predicted DCR, an early clinical activity indicator, although they did not predict OS or PFS. These outcomes pertain to a NSCLC subset and might not apply to other tumor types. Tumoral AXL expression did not linearly correlate with EnaV activity in a patient-derived soft tissue sarcoma model3, possibly due to AXL expression in both tumor cells and myeloid-derived suppressor cells, or the presence of tumor-intrinsic MMAE resistance4, implying the AXL expression and EnaV activity relationship is contingent on the tumor microenvironment5. Further validation in larger cohorts is planned. 1 Davatzikos C, et al. J Med. Imaging 2018 2 Horng H, et al. Sci. Reports 2022 3 Van Renterghem B, et al. Int J Mol Sci. 2022 4 Holtzhausen A, et al. Cancer Immunol Res 2019 5 Boshuizen J, et al. Cancer Res. 2021 Citation Format: Andrew W. Chen, Lauren K. Brady, Eric A. Cohen, Walter C. Mankowski, Leonid Roshkovan, Mohammed Qutaish, Sharyn Katz, Patricia G. Castro, Nora Pencheva, Maria Jure-Kunkel, Brandon W. Higgs, Despina Kontos. Evaluating the potential of radiomics features to predict outcome to an antibody-drug conjugate therapy in non-small cell lung cancer patients [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2587.

Abstract 6247: Comprehensive characterization of cancers with epithelial mesenchymal transition phenotype through long-read sequencing

Abstract Introduction: Pancreatic ductal adenocarcinoma (PDAC) and linitis-plastica type gastric cancer (STAD) are the most representative refractory cancers even in the era of molecularly targeted therapy and immunotherapy. Extensive characterization of these tumors, especially regarding the genomic and epigenomic characteristics of pure cancer cells, has been hampered by the high stromal content in the primary tumor samples. Here we took advantage of cancer cells purified from malignant ascites and established corresponding cell lines to overcome the limitations associated with the analysis and to discover their vulnerabilities. Methods: We developed cell lines from malignant ascitic fluid samples obtained from 60 patients with PDAC and 59 patients with linitis-plastica type STAD. We performed comprehensive multi-omic analyses of these cell lines using long-read whole-genome sequencing (WGS) for complex somatic genomic alterations and for methylation aberration. We also conducted long-read RNA sequencing (iso-seq) for full-length isoform exploration, and chromatin immunoprecipitation coupled with sequencing (ChIP-seq) for enhancer analysis. Complementary short-read whole-genome sequencing and RNA sequencing were also performed. Results: Comprehensive transcriptome analyses identified a cluster with high epithelial-mesenchymal transition (EMT) characteristics, consisting of the majority of the PDAC cell lines and a half of the STAD cell lines. ChIP-seq analysis against H3K27ac revealed the distinct transcriptional circuit that generated EMT cluster-specific super-enhancers governed by SMAD3, independent of cancer type. Most of the transcriptional activity was controlled by the gene methylation. We also identified the differential expression levels of the splicing factors between the EMT-activated and the non-activated groups. Comprehensive isoform analysis by iso-seq revealed that the isoform usage was completely different between the EMT group and the other, especially in genes of the actomyosin pathway which confer characteristic features of the EMT cluster in the cell morphology and cell motility. Using the CHM13 reference genome, long-read WGS analyses revealed the recurrent structural alterations specific to the EMT group such as those involving CDKN2A. We also identified PDAC-specific structural alterations. Conclusion: We identified the distinct molecular characteristics of the EMT-activated group in PDAC and linitis-plastica type STAD which may confer the therapeutic resistance. Our findings provide an interesting insight for the further innovation of the treatment strategy for these intractable cancers. Citation Format: Yosuke Tanaka, Hiroyuki Mano. Comprehensive characterization of cancers with epithelial mesenchymal transition phenotype through long-read sequencing [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6247.

Особенности проектирования кластеров как больших организационно-экономических систем

The need for accelerated reindustrialization of industry highlights the task of modernization based on modern principles of the procedure for designing new enterprises, holdings, clusters and other complex economic structures, united by a category of large organizational and economic systems. The problem is that if the development of a promising image of an industrial enterprise is carried out using an established technology (analysis of the initial data – development strategy development – building a business model – development of projects of production and social subsystems, as well as management subsystems – preparation of a business plan), then the procedure for designing an industrial cluster is not yet clearly regulated. Only the requirements for industrial clusters applying for state support for their development have been identified. The absence of such regulations leads to a situation in which some of the already created industrial clusters cease to exist due to miscalculations made during their formation and the organization of the initial stage of operation. The features of organizational design of industrial clusters – a promising form of regional concentration of production are consiered. The article discloses the features of industrial clusters and their differences from industrial enterprises from the point of view of design, and also proposes a procedure for performing this work, which includes an original set of actions that ensures the achievement of the goals of the participants in this association. Its specificity is associated with the cooperation of legally independent economic entities from various spheres of the economy, interested in obtaining an additional effect from the coordination of their activities in the region, on the terms of informal unification without losing the freedom to choose their areas of activity, but within the framework of coordinating the joint achievement of common goals.

Abstract 2310: Integrating real-world histopathological and clinicogenomic data from 1799 lung cancer patients by applying unsupervised deep learning

Abstract Hematoxylin and eosin (H&E) stained histologic sections contain invaluable information that remains largely untapped because of its complexity. To this end, AI applications employing deep learning (DL) can facilitate the translation of image data to enable human interpretation and yield novel oncological insights that would have otherwise remained imperceptible. DL-based methodologies are multimodal, capable of integrating imaging with clinicogenomic data to furnish a more holistic perspective and affording more accurate predictions in oncology. Here, we developed an unsupervised DL workflow to analyze 1,799 H&E images of lung cancer (NSCLC n = 951; SCLC n = 50; others n = 798) incorporating comprehensive patient-level clinical data (electronic health records [ConcertAI]) integrated with genomics (WES and RNA-seq [Caris Labs]). There are three steps in our approach: (1) image preprocessing and filtering, yielding > 30 million image patches; (2) utilizing pretrained SimCLR models from 57 public oncology histopathology datasets with ResNet-18 as a backbone structure to extract 512-dimensional-feature vectors for each patch; (3) using three unsupervised clustering methods (kmeans, DBSCAN, Leiden clustering) to cluster patches and selected Leiden clustering. We identified 635 primary imaging clusters using an elbow method and generated an image feature matrix by calculating correlations between each patch and cluster centroids; these were aggregated and mapped back to source slides. In this proof-of-concept, distinct image feature patterns characterized SCLC and NSCLC samples. For SCLC, one of the salient features was the presence of hemorrhage, which may be associated with higher rates of fine-needle aspiration biopsy procedure for SCLC compared with NSCLC which was confirmed in the EHR data (p = 0.032). Derived morphological clusters were correlated with tumor-immune genomic features (Tumor Mutational Burden [TMB], Immunologic Constant of Rejection [ICR], and Miracle scores1) serving as predictors of response to immune-checkpoint inhibitor therapy. By applying linear models, we detected 11, 96 and 249 significantly associated imaging clusters, respectively, highly enriched with immune cells e.g., plasma cells, macrophages, lymphocytes, and supporting an infiltrated and inflamed tumor-immune microenvironment. In summary, a multimodal, unsupervised deep learning workflow combining H&E imaging with clinicogenomic data was developed to identify histologic feature clusters associated with well-established tumor-immune genomic signatures of NSCLC immune infiltration and molecular phenotypes. These studies demonstrate enormous potential to yield histopathological and translational insights in NSCLC and SCLC that can empower clinicians to make better therapeutic response predictions. Citation Format: Si Wu, Yujie Zhao, Hugo Luo, Kevin Kolahi, Thanh Bui, Xu Shi, Aditee Shrotre, Alexander Liede, Xi Zhao, Josue Samayoa, Weilong Zhao. Integrating real-world histopathological and clinicogenomic data from 1799 lung cancer patients by applying unsupervised deep learning [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2310.

Abstract 5479: Spatially resolved molecular features in the tumor-immune microenvironment of germline BRCA1/2 mutated versus BRCA wild type high grade serous ovarian cancers

Abstract High-grade serous ovarian cancer (HGSC) is the most common and aggressive subtype of epithelial ovarian cancer. Importantly, germline mutations in BRCA1/2 give rise to HGSC. Because most HGSC are diagnosed at stage III/IV, outcomes are poor and more effective therapies are needed. Although immune checkpoint inhibitors show promise for a number of cancer types, efficacy has been variable in HGSC. To help provide new clues to the molecular features of HGSC, we set out to elucidate the tumor microenvironment of HGSC using spatial transcriptomics approaches. Specifically, we applied 10x Genomics Visium Spatial Transcriptomics (ST) technology to 48 samples representing primary HGSC tumors from 24 patients, including 15 BRCA1/2 mutated cases and 9 BRCA non-mutated cases. In addition to individual sample analysis, we normalized data from 64,784 total features across all sections and performed an integrated cluster analysis to get an overall picture of the cellular architecture of HGSC, but to also directly investigate differences in the tumor microenvironment based on BRCA mutation status. The resulting 17 integrated clusters were subjected to molecular annotation using GSEA, ESTIMATE, and CIBERSORTx, which revealed feature clusters representing important cell types defined by marker genes. Analysis of the entire dataset identified a primary tumor cluster defined by the expression of known ovarian cancer genes including WFDC2 and “don’t eat me” marker genes CD24 and CD47. However, the predominant tumor cell population in BRCA mutated cases was marked by the expression of MUC1, EPCAM, CLDN3, and CD24, while BRCA non-mutated tumor cell cluster was marked by EFNA1, ID1, LSR, and CD9. Importantly, both groups revealed the presence of a specialized monocyte derived macrophage cluster defined by the expression of LYZ and genes involved in cholesterol metabolism including APOE and APOC1. These lipid-associated macrophages are M2 polarized and shown to be associated with poor outcome. Additional analyses are underway to determine spatial dependencies of various tumor and immunological cell compartments. These data provide novel insights into the HGSC tumor microenvironment that may influence HGSC’s vulnerability to immunological therapies. Citation Format: Jing Qian, Lynda D. Roman, Rania Bassiouni, Seeta Rajpara, Monica Neuman, Varun Khetan, David W. Craig, Joseph W. Carlson, John D. Carpten. Spatially resolved molecular features in the tumor-immune microenvironment of germline BRCA1/2 mutated versus BRCA wild type high grade serous ovarian cancers [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 5479.

Galaxy assembly revealed by globular clusters

Many observable properties of globular clusters (GCs) provide valuable insights for unveiling the hierarchical assembly of their host galaxy. For the Milky Way (MW) in particular, GCs from different accreted satellite galaxies show distinct chemical, spatial, kinematic, and age distributions. Here we examine such clustering features for model GC populations in simulated galaxies, which are carefully selected to match various observational constraints of the MW assembly. We evaluate several widely used clustering, dimensionality reduction, and supervised classification methods on these model GCs, using 10 properties that are observable in the MW. We can categorize in-situ and ex-situ formed GCs with about 90% accuracy, based solely on their clustering features in these 10 variables. The methods are also effective in distinguishing the last major merger in MW analogs with similar accuracy. Although challenging, we still find it possible to identify one, and only one, additional smaller satellite. We develop a new technique to classify the progenitors of MW GCs by combining several methods and weighting them by the validated accuracy. According to this technique, about 60% of GCs belong to the in-situ group, 20% are associated with the Gaia-Sausage/Enceladus event, and 10% are associated with the Sagittarius dwarf galaxy. The remaining 10% of GCs cannot be reliably associated with any single accretion event.

Time series prediction for production quality in a machining system using spatial-temporal multi-task graph learning

The quality of multiple-machining-feature parts in batch computer numerical control machining systems is affected by complex spatial-temporal relationships, making accurate prediction difficult. Traditional quality prediction methods are difficult to cope with the challenges posed by the time impact effects generated by batch production sequences and the spatial impact effects generated by the adjacency of machining features of thin-walled parts. Hence, the time series prediction for production quality in a machining system using spatial-temporal multi-task graph learning is proposed to analyze the spatial-temporal impact relationship of machining quality and achieve accurate quality prediction. Specifically, a graph learning-based framework to capture the coupled spatial-temporal effects is designed to represent the spatial topological and temporal relationship of the machining features of batch parts. A novel global-local multi-task spatial-temporal graph learning framework is constructed to simultaneously predict the time series quality of each machining feature at different spatial positions. The global multi-task model aims to extract spatial-temporal coupling data features (i.e., input parameters for machining features at different positions of batch parts) for machining feature clusters based on gated networks and spatial-temporal graph convolution with multi-head attention. Then, the local multi-task model for each feature cluster (round holes, rectangular grooves, etc.) performs individual time series quality prediction regarding multiple quality indicators. Experimental results on a production dataset of a computer numerical control machining system for thin-walled parts show that the proposed method is superior to the conventional models such as a simple spatial-temporal graph convolution method (3.0%, 18.3% and 11.0% improvement in R2, MAE and RMSE) or the model without graph approach (6.5%, 29.8% and 19.8% improvement in R2, MAE and RMSE).

Topic Clusters of Scientific Organizations of Moscow Healthcare Department: Distribution and Leaders

Introduction. Studying the evolution and life cycle of topics that form a particular field of knowledge is important in order to predict certain trends in the development of scientific research in any scientific field (including medicine), as these topics are crucial in identifying and developing new directions. The SciVal topic cluster system is one of the effective, appealing, and promising models for topic-based article classification. It is utilized on the SciVal analytical platform using data from the Scopus international scientific citation database.
 Materials and methods. Data collection and uploading were carried out as of January 24, 2023. The time frame of the study was four years, from 2019 to 2022. The goal was to examine a set of medical publications for 15 organizations (4 research institutes and 11 scientific and practical centers) subordinate to Moscow Healthcare Department.
 Results and discussion. Scientific organizations were ranked according to six scientometric indicators that describe the share distribution and quantitative characteristics of medical topic clusters. The study determined the leaders among all the 15 organizations as well as a top three among research institutes and scientific and practical centers. Topic clusters were analyzed based on the maximum and minimum values in the number of publications and citation indexes across 15 organizations and separately for two groups (research institutes and scientific and practical centers).
 Conclusion. One of the primary paths of future development and implementation of the methods presented in this paper appears to be the use of matrix analysis to identify and determine the key features of both individual subject clusters and joint thematic clusters. With the use of matrix analysis, we will be able to establish the most extensive, referenced, and relevant topic clusters based on the quantitative and equity ratios that we calculate not only between the topic clusters themselves but also between organizations that have publications in these clusters.

Sentiment analysis incorporating convolutional neural network into hidden Markov model

AbstractThe analysis of sentiments and mining of opinions have become more and more important in years because of the development of social media technologies. The methods that utilize natural language processing and lexicon‐based sentiment analysis techniques to analyze people's opinions in texts require the proper extraction of sentiment words to ensure accuracy. The current issue is tackled with a novel perspective in this paper by introducing a hybrid sentiment analysis technique. This technique brings together Convolutional Neural Network (CNN) and Hidden Markov Models (HMMs), to accurately categorize text data and pinpoint feelings. The proposed method involves 1D convolutional‐layer CNN to extract hidden features from comments and applying HMMs on a feature‐sentence matrix, allowing for the utilization of word sequences in extracting opinions. The method effectively captures diverse text patterns by extracting a range of features from texts using CNN. Text patterns are learned using text HMM by calculating the probabilities between sequences of feature vectors and clustering feature vectors. The paper's experimental evaluation employs benchmark datasets such as CR, MR, Subj, and SST2, demonstrating that the proposed method surpasses existing sentiment analysis techniques and traditional HMMs. One of its strengths is to analyze a range of text patterns and identify crucial features that recognize the emotion of different pieces of a sentence. Additionally, the research findings highlight the improved performance of sentiment analysis tasks through the strategic use of zero padding in conjunction with the masking technique.

Domain-aware triplet loss in domain generalization

Despite the considerable advances in deep learning for object recognition, there are still several factors that hinder the performance of deep learning models. One of these factors is domain shift, which occurs due to variations in the distribution of the testing and training data. This paper addresses the issue of compact feature clustering in domain generalization, with the aim of optimizing the embedding space from multi-domain data. Specifically, we propose a domain-aware triplet loss for domain generalization, which not only facilitates clustering of similar semantic features but also disperses features that arise from the domain. Unlike previous methods that focus on aligning distributions, our algorithm disperses domain information in the embedding space. Our approach is based on the assumption that embedding features can be clustered based on domain information, which is supported mathematically and empirically in this paper.Furthermore, in our investigation of feature clustering in domain generalization, we observe that the factors that influence the convergence of metric learning loss in domain generalization are more significant than the pre-defined domains. To address this issue, we utilize two methods to normalize the embedding space and reduce the internal covariate shift of the embedding features. Our ablation study illustrates the effectiveness of our algorithm. Additionally, our experiments on benchmark datasets, including PACS, VLCS, and Office-Home, demonstrate that our method outperforms related approaches that focus on domain discrepancy. Notably, our results on RegnetY-16GF are substantially better than state-of-the-art methods on the benchmark datasets. Our code is available at https://github.com/workerbcd/DCT.

SERS analysis, DFT, and solution effects regarding the structural and optical characteristics of folic acid biomolecule adsorbed on a Cu3 metal cluster

The optical characteristics of folic acid (ABP) and metal clusters of copper (Cu3) at various locations were investigated by means of density functional theory (DFT) computations. Mulliken charge analysis and molecular electrostatic potential (MEP) surface show how charge moves from Cu3 to ABP through the various groups. The peak in the UV–Vis spectra of ABP-Cu3 is caused by bonding and anti-bonding orbitals. In both vacuum and aqueous conditions, the polarizability values of ABP-Cu3 cluster are significantly higher than those of pure ABP, indicating a possible enhancement of the nonlinear optical (NLO) effect. Our research investigates the possibility of using ABP adsorbed metal clusters for NLO materials. Surface enhanced Raman scattering (SERS) in the ABP adsorbed metal clusters enhances the vibrational modes of ABP. Adsorption energies are found to be in the range −17.08 to −58.52 kcal/mol in vacuum and −53.34 to −93.44 kcal/mol in aqueous medium for the different configurations for ABP-Cu3. It indicates that metal clusters adsorbed by ABP are stable in the aqueous media. Experimental IR and UV–Vis of ABP is in agreement with theoretically predicted ones.

OralExplorer: a web server for exploring the mechanisms of oral inflammatory diseases.

Oral inflammatory diseases are localized infectious diseases primarily caused by oral pathogens with the potential for serious systemic complications. However, publicly available datasets for these diseases are underutilized. To address this issue, a web tool called OralExplorer was developed. This tool integrates the available data and provides comprehensive online bioinformatic analysis. Human oral inflammatory disease-related datasets were obtained from the GEO database and normalized using a standardized process. Transcriptome data were then subjected to differential gene expression analysis, immune infiltration analysis, correlation analysis, pathway enrichment analysis, and visualization. The single-cell sequencing data was visualized as cluster plot, feature plot, and heatmaps. The web platform was primarily built using Shiny. The biomarkers identified in OralExplorer were validated using local clinical samples through qPCR and IHC. A total of 35 human oral inflammatory disease-related datasets, covering 6 main disease types and 901 samples, were included in the study to identify potential molecular signatures of the mechanisms of oral diseases. OralExplorer consists of 5 main analysis modules (differential gene expression analysis, immune infiltration analysis, correlation analysis, pathway enrichment analysis and single-cell analysis), with multiple visualization options. The platform offers a simple and intuitive interface, high-quality images for visualization, and detailed analysis results tables for easy access by users. Six markers (IL1β, SRGN, CXCR1, FGR, ARHGEF2, and PTAFR) were identified by OralExplorer. qPCR- and IHC-based experimental validation showed significantly higher levels of these genes in the periodontitis group. OralExplorer is a comprehensive analytical platform for oral inflammatory diseases. It allows users to interactively explore the molecular mechanisms underlying the action and regression of these diseases. It also aids dental researchers in unlocking the potential value of transcriptomics data related to oral diseases. OralExplorer can be accessed at https://smuonco.shinyapps.io/OralExplorer/ (Alternate URL: http://robinl-lab.com/OralExplorer ).

Structure-guided feature and cluster contrastive learning for multi-view clustering

Multi-view clustering (MVC) technology performs unsupervised clustering on data collected from multiple sources, and has received intense attention in recent years. However, most existing MVC methods fail to consider retaining view-specific information when learning multi-view consistent representations. Besides, the feature and cluster structures of multi-view data cannot be fully leveraged in clustering. In this paper, we propose a structure-guided feature and cluster contrastive learning (SGFCC) for multi-view clustering. Specifically, SGFCC utilizes autoencoders to achieve view-specific information reconstruction in feature space, and extracts high-level features for multi-view consistent representation learning to eliminate the effects of view-specific information and noise on consistent representation. To fully capture the similar clustering structure of high-level features and semantic features of samples across different views, we adopt a structure-guided feature-level and cluster-level contrastive learning strategy in our SGFCC model. Furthermore, we design a clustering layer to explore the cluster structure of high-level features. Different from most existing MVC methods, our method applies a non-fusion scheme that aggregates the semantic information of all views to obtain the final semantic labels. Extensive experiments on public datasets demonstrate that our method outperforms other competitors in clustering tasks.

Prevalence and Molecular Characterization of Vancomycin Variable Enterococcus faecium Isolated From Clinical Specimens.

Vancomycin variable Enterococcus (VVE) bacteria are phenotypically susceptible to vancomycin, but they harbor the vanA gene. We aimed to ascertain the prevalence of VVE among clinically isolated vancomycin-susceptible Enterococcus faecium (VSE) isolates, as well as elucidate the molecular characteristics of the vanA gene cluster within these isolates. Notably, we investigated the prevalence and structure of the vanA gene cluster of VVE. Between June 2021 and May 2022, we collected consecutive, non-duplicated vancomycin-susceptible Enterococcus faecium (VSE) samples. Real-time PCR was performed to detect the presence of vanA, vanB, and vanC. Overlapping PCR with sequencing and whole-genome sequencing were performed for structural analysis. Sequence types (STs) were determined by multilocus sequence typing. Exposure testing was performed to assess the ability of the isolates to acquire vancomycin resistance. Among 282 VSE isolates tested, 20 isolates (7.1%) were VVE. Among them, 17 isolates had partial deletions in the IS1216 or IS1542 sequences in vanS (N=10), vanR (N=5), or vanH (N=2). All VVE isolates belonged to the CC17 complex and comprised five STs, namely ST17 (40.0%), ST1421 (25.0%), ST80 (25.0%), ST787 (5.0%), and ST981 (5.0%). Most isolates were related to three hospital-associated clones (ST17, ST1421, and ST80). After vancomycin exposure, 18 of the 20 VVEs acquired vancomycin resistance. Considering the high reversion rate, detecting VVE by screening VSE for vanA is critical for appropriate treatment and infection control.

The Distribution and Accessibility of Elements of Tourism in Historic and Cultural Cities

Historic urban areas are the foundations of urban development. Due to rapid urbanization, the sustainable development of historic urban areas has become challenging for many cities. Elements of tourism and tourism service facilities play an important role in the sustainable development of historic areas. This study analyzed policies related to tourism in Panguifang and Meixian districts in Meizhou, Guangdong, China. Kernel density estimation was used to study the clustering characteristics of tourism elements through point of interest (POI) data, while space syntax was used to study the accessibility of roads. In addition, the Pearson correlation coefficient and regression were used to analyze the correlation between the elements and accessibility. The results show the following: (1) the overall number of tourism elements was high on the western side of the districts and low on the eastern one, and the elements were predominantly distributed along the main transportation arteries; (2) according to the integration degree and depth value, the western side was easier to access than the eastern one; and (3) the depth value of the area negatively correlated with kernel density, while the degree of integration positively correlated with it. Based on the results, the study put forward measures for optimizing the elements of tourism in Meizhou’s historic urban area to improve cultural tourism and emphasize the importance of the elements.