Clustering Output Research Articles

The division of PM2.5-O3 composite airborne pollution across China based on spatiotemporal clustering

With the rapid increase of ground-level ozone concentrations, the comprehensive management of PM2.5-O3 composite air pollution has become one of the most pressing environmental concerns nowadays. However, due to the lack of national divisions, regional integrative management of PM2.5-O3 composite air pollution remains highly challenging. To fill this gap, we employed and adapted a repeated-bisection model to conduct spatiotemporal clustering of PM2.5-O3 composite airborne pollution across China based on multi-year airborne pollutant data in 364 cities. Specifically, two strategies were experimented: the spatiotemporal clustering of daily PM2.5/O3 and the spatiotemporal clustering of daily PM2.5 and O3 concentrations. Despite some differences, the clustering outputs from both strategies achieved a self-aggregation effect, indicating that cities with similar spatiotemporal patterns of simultaneous PM2.5 and O3 variations were usually located closely. This phenomenon suggests the necessity and feasibility of regional integrative management of composite airborne pollution. According to accuracy assessment based on Geographical Detector, both strategies achieved relatively satisfactory outputs. Specifically, the spatiotemporal clustering based on daily PM2.5 and O3 concentrations achieved a slightly better output, suggesting PM2.5/O3 cannot fully explain the complicated and uncertain PM2.5-O3 association. Based on the clustering output, we divided seven divisions of PM2.5-O3 composite airborne pollution across China. This research provides important decision support for conducting regional integrative management of composite airborne pollution. The framework of two-variable-oriented spatiotemporal clustering sheds useful light on the comprehensive management of multiple and mutually-interacting environmental issues.

Integrating spatial dependence into functional clustering of NDVI in the Ecuadorian Andes

AbstractSpatial dependence into environmental data is an influential criterion in clustering processes, as the resulting clustering outputs depend very much upon such spatial structure. As classical methods do not take spatial dependence in consideration, the inclusion of this structure produces unexpected but more realistic results and clusters of curves that may not be similar in shape or behavior. In this paper, clustering is made using the KMSCFD algorithm for spatially correlated functional data. The methodology was developed through weighting the distance matrix between the curves with the trace‐variogram calculated with the coefficients of the basis functions resulting from a data smoothing operation. For the validation of the method, a number of simulated scenarios were tested together with an application to Normalized Difference Vegetation Index data derived from a high elevation ecosystem in the Ecuadorian Andes. Quality indices are implemented to obtain the appropriate number of clusters. The analysis showed five different regions that were latitudinally distributed.

A novel heterogenous ensemble theory for symmetric 5G cells segmentation: Intelligent RAN analytics

MNOs are investing more in 5G, rolling out sites in urban and specific rural areas. Meanwhile, it remains imperative to consistently maintain the network performance above a certain threshold for optimal user experience. Symmetric network cells characterized by parallel attributes in terms of capacity and coverage are instrumental in planning, optimization, and resource allocation. However, the variation in environmental factors introduces divergence in network cells' behavior, symmetric or not. Therefore, the need arises for intelligent analytic processes within the RAN system to categorize symmetric cells based on their performance and behavior. Intelligent optimization and analytics in 5G rely on the accurate and automated identification of cells exhibiting symmetric behavior, enabling bulk optimization operations. In this paper, we develop and assess a clustering approach using a heterogenous ensemble method to group 5G cells based on their key performance attributes to facilitate network optimization tasks. The approach involves a synergistic integration of K-means and hierarchical clustering algorithms, enabling dynamic segmentation of cells based on their performance behavior. Leveraging the clustering output, we train an XGBoost classifier, paving the way for a comprehensive analytics framework and problematic or poor-performing cells’ detection. We apply the study model to real-world 5G RAN metrics and evaluate the proposed method in terms of clustering accuracy and convergence. The study output showcases the efficacity of the heterogenous ensemble approach compared to individual clustering algorithms, providing a valuable baseline for network performance enhancement. With such a dynamic approach for analyzing 5G new radio (NR) performance, MNOs can move toward intelligent and self-aware networks, making informed decisions regarding resource allocation and coverage optimization.

Multivariate Time Series Clustering of Groundwater Quality Data to Develop Data-Driven Monitoring Strategies in a Historically Contaminated Urban Area

As groundwater quality monitoring networks have been expanded over the last decades, significant time series are now available. Therefore, a scientific effort is needed to explore innovative techniques for groundwater quality time series exploitation. In this work, time series exploratory analysis and time series cluster analysis are applied to groundwater contamination data with the aim of developing data-driven monitoring strategies. The study area is an urban area characterized by several superimposing historical contamination sources and a complex hydrogeological setting. A multivariate time series cluster analysis was performed on PCE and TCE concentrations data over a 10 years time span. The time series clustering was performed based on the Dynamic Time Warping method. The results of the clustering identified 3 clusters associated with diffuse background contamination and 7 clusters associated with local hotspots, characterized by specific time profiles. Similarly, a univariate time series cluster analysis was applied to Cr(VI) data, identifying 3 background clusters and 7 hotspots, including 4 singletons. The clustering outputs provided the basis for the implementation of data-driven monitoring strategies and early warning systems. For the clusters associated with diffuse background contaminations and those with constant trends, trigger levels were calculated with the 95° percentile, constituting future threshold values for early warnings. For the clusters with pluriannual trends, either oscillatory or monotonous, specific monitoring strategies were proposed based on trends’ directions. Results show that the spatio-temporal overview of the data variability obtained from the time series cluster analysis helped to extract relevant information from the data while neglecting measurements noise and uncertainty, supporting the implementation of a more efficient groundwater quality monitoring.

Power Generation Scheduling for a Hydro-Wind-Solar Hybrid System: A Systematic Survey and Prospect

In the past two decades, clean energy such as hydro, wind, and solar power has achieved significant development under the “green recovery” global goal, and it may become the key method for countries to realize a low-carbon energy system. Here, the development of renewable energy power generation, the typical hydro-wind-photovoltaic complementary practical project, is summarized, and some key problems in complementary systems such as the description and prediction of the power generation law in large-scale stations, risk management, and coordinated operation are analyzed. In terms of these problems, this paper systematically summarizes the research methods and characteristics of a hydro-wind-solar hybrid system and expounds upon the technical realization process from the prediction and description of wind and solar power station cluster output, the risks brought about by large-scale renewable energy grid-connected operation, and the long-term and short-term coordination modeling and resolution thoughts on the hydro-wind-solar hybrid system in cluster mode. Finally, based on the aforementioned analysis, the existing research gaps are discussed from the standpoints of generation forecast, risk management, and cluster scheduling, and the future work outlook is presented accordingly. A hybrid system that combines hydro, wind, and solar energy is emerging as a way to make up for each other’s shortcomings and will be a fruitful area of study in the future.

A Hybrid Intelligent Clustering Model for Tackling Incomplete Mixed Data Using Heuristic Algorithm with Artificial Intelligence

Recently, in various information areas including image processing and social networks, data clustering has been vastly utilized. Thus, the clustering task is taken as successful while imputing the quality of uncertain values, whereas the traditional methods observe only poor performance. To enhance the performance, the hybrid intelligent-based clustering model is proposed by the adaptive concept. Initially, the data is collected from benchmark datasets that are composed of incomplete mixed data. The first level of clustering is processed by optimized K-Means Clustering (KMC) to acquire the optimal centroid, in which the centroid is optimized by the Adaptive Probability-assisted Water Strider Algorithm (AP-WSA). Similarly, the second clustering is done through K-Correlation-based Clustering (KCC) to obtain the clustered output. Thus, the final clustering process is named as Hybrid K-Means K-Correlation (HKMKC) clustering algorithm. Finally, the experimental results are validated with the performance measures and estimated with the traditional methods. Through the result analysis, the Average Silhouette Coefficient (AS) of the designed AP-WSA-HKMKC is 50%, 28%, 16%, and 5.45% progressed than KMC, FCM, KCC, and K-median clustering methods for dataset 2 at 15% miss rate. Thus, the designed hybrid clustering ensures the efficiency of handling incomplete data regarding accurate clustered outputs.

A Star-Identification Algorithm Based on Global Multi-Triangle Voting

A star-identification algorithm aimed at identifying imaged stars in a “lost in space” scene, named the global multi-triangle voting algorithm (GMTV), is presented in this paper. There are two core parts included in the proposed algorithm: in the initial match part, triangle feature units are treated as vote units to find the initial match relationship via matching vote units and counting the vote number of each catalog star. During this step, the principal component analysis (PCA) method is implemented to reduce feature dimensions, and a two-dimension lookup table and fuzzy match strategy are utilized to promote database searching efficiency and noise tolerance. After acquiring the initial match results, a verification part is implemented to filter potential errors from initial candidates by the largest cluster method and output the final identification results. The proposed algorithm achieves a 98.6% identification rate with 2.0 pixels position noise and exhibits more robustness to position noise, magnitude noise, and false stars of different levels than the two reference algorithms used in simulations. In addition, our algorithm’s real-time performance is better than reference algorithms, but it requires a larger database.

Graph Convolutional Network Combined with Semantic Feature Guidance for Deep Clustering

The performances of semisupervised clustering for unlabeled data are often superior to those of unsupervised learning, which indicates that semantic information attached to clusters can significantly improve feature representation capability. In a graph convolutional network (GCN), each node contains information about itself and its neighbors that is beneficial to common and unique features among samples. Combining these findings, we propose a deep clustering method based on GCN and semantic feature guidance (GFDC) in which a deep convolutional network is used as a feature generator, and a GCN with a softmax layer performs clustering assignment. First, the diversity and amount of input information are enhanced to generate highly useful representations for downstream tasks. Subsequently, the topological graph is constructed to express the spatial relationship of features. For a pair of datasets, feature correspondence constraints are used to regularize clustering loss, and clustering outputs are iteratively optimized. Three external evaluation indicators, i.e., clustering accuracy, normalized mutual information, and the adjusted Rand index, and an internal indicator, i.e., the Davidson-Bouldin index (DBI), are employed to evaluate clustering performances. Experimental results on eight public datasets show that the GFDC algorithm is significantly better than the majority of competitive clustering methods, i.e., its clustering accuracy is 20% higher than the best clustering method on the United States Postal Service dataset. The GFDC algorithm also has the highest accuracy on the smaller Amazon and Caltech datasets. Moreover, DBI indicates the dispersion of cluster distribution and compactness within the cluster.

Bi-level clustering of vehicle trajectories for path choice set and its nested structure identification

Path choice set identification is essential for route choice modelling and travel behaviour studies. Recent advancements in data collection techniques have gained attention towards a data-driven choice set identification process. However, empirical vehicle trajectory datasets result in several path observations compared to the traditional algorithms, complicating the route choice modelling process. This study proposes a bi-level vehicle trajectory clustering framework where the output of the upper-level clustering provides a representative path choice set for simple/mixed logit modelling (MNL), whereas the lower-level clustering provides a nested or cross-nested representation of the paths based on hard and soft clustering, respectively.As proof of concept, the proposed methodology is applied on real Bluetooth-based trajectories from Brisbane, where 62 unique paths were observed from a one-year trajectory data for an origin–destination pair. The results of the MNL model for the representative paths provide desirable magnitude with negative coefficients for the distance and travel time path attributes. Further, the results of the (cross) nested modelling appropriately identified the (cross) nested structure for the path choice set.

The importance of taking ART appropriately in children and adolescents with HIV-1 to reach the highest capacity of immune function later in life.

Current antiretroviral therapy (ART) guidelines recommend treating all children with HIV-1 infection. This has changed from the broader use of ART to treat children to improve morbidity and minimise mortality. However, prior to current recommendations, not everyone with HIV-1 received timely treatment. What happens to the paediatric immune system when HIV-1 replication is not appropriately supressed remains unclear. 11 samples from adolescents with HIV-1 on ART and uninfected controls in the UK, aged 12–25 years, were examined; overall, adolescents with CD4+ counts > 500/μl and a viral load < 50 copies/ml were compared with adolescents with CD4+ counts < 500/μl and a viral load > 50 copies/ml at time of sampling. Measurements of thymic output were combined with high throughput next generation sequencing and bioinformatics to systematically organize CD4+ and CD8+ T cell receptor (TCR) repertoires. TCR repertoire diversity, clonal expansions, TCR sequence sharing, and formation of TCR clusters in HIV-1 infected adolescents with successful HIV-1 suppression were compared to adolescents with ineffective HIV-1 suppression. Thymic output and CD4+ T cell numbers were decreased in HIV-1 infected adolescents with poor HIV-1 suppression. A strong homeostatic TCR response, driven by the decreased CD4+ T cell compartment and reduced thymic output was observed in the virally uncontrolled HIV-1-infected adolescents. Formation of abundant robust TCR clusters and structurally related TCRs were found in the adolescents with effective HIV-1 suppression. Numerous CD4+ T cell numbers in the virally controlled adolescents emphasize the importance of high thymic output and formation of robust TCR clusters in the maintenance of HIV-1 suppression. While the profound capacity for immune recovery in children may allow better opportunity to deal with immunological stress, when ART is taken appropriately, this study demonstrates new insights into the unique paediatric immune system and the immunological changes when HIV-1 replication is ongoing.

Transient Stability Evaluation Criterion of Multi-Wind Farms Integrated Power System

This letter proposes a novel transient stability criterion for power systems with multi-wind farms, which is deduced from power flow and inertia distribution. The variations of acceleration and deceleration power with and without wind integration, instead of the acceleration and deceleration areas themselves, are calculated, in which the power angle swing direction, the transient stability control of wind turbines, the position of wind farms access, and the unbalanced output of synchronous clusters are also considered. The simulation results show the effectiveness of the proposed criterion to distinguish transient stability.

Implementasi Algoritma K-Means Clustering Dalam Penilaian Kedisiplinan Siswa

Education has a very important role for students, not just potential but noble character in the form of discipline, therefore it is necessary to group each school based on student discipline. Implementing a clustering system with the K-means method which is used to classify and determine the value of student discipline which produces a clustering output of student discipline that is beneficial for the school to prevent students from misbehaving early on. Analysis of data needs used in this study in the form of primary data obtained from a questionnaire given to students. The attributes used are presence, neatness and behavior. Student discipline assessment can be carried out using the k-means clustering method. This study applies the K-Means clustering algorithm method using Microsoft Excel 2013 and Orange which performs the data mining process. The results of the research implementation of the k-means clustering algorithm in student disciplines are divided into three clusters. From a sample of 133 students, 41 students were included in cluster one (C1), then 33 students were included in the second cluster (C2), and 59 students were included in cluster three (C3). The results of grouping the level of student discipline using the k-means method can be used as a reference or assessment of the discipline of each student.

Improved Regional Frequency Analysis of rainfall data

Rainfall are subject to local orography features and their intensities can be highly variable. In this context, identifying climatically coherent regions can greatly help interpreting rainfall patterns and improve the inference of return levels. In practice, partitioning a region of interest into homogeneous sub-regions is a delicate statistical task, especially in regards to modeling heavy rainfall features.In this work, our main goal is to propose and study a fast and efficient clustering algorithm. Compared to classical regional frequency analysis techniques, a key aspect is that our algorithm does not rely on the a priori choice of covariates. The proposed numerical scheme is only based on the precipitation dataset at hand, including low, moderate and heavy rainfall. In terms of inference, our approach builds on the easy-to-compute and reliable probability weighted moments commonly used in hydrology. While being in compliance with extreme value theory, we do not impose a parametric form on rainfall distributions and, neither thresholding nor block maxima steps are required in our proposed approach. By construction, our clustering method preserves the scale invariance principle of any classical regional frequency analysis.The performance of our clustering algorithm is assessed on a detailed experimental design based on the extended Generalized Pareto distribution.Sensitivity to the number of clusters is carefully analyzed.We apply our clustering algorithm on Switzerland daily precipitation measured at 191 sites. The found homogeneous regions are consistent with local orography and our approach outperforms the classical regional frequency analysis based on normalized elevation and coordinates as covariates. To complete our analysis of Swiss rainfall, we propose three models based on our clustering outputs. A comparison between our local, semi-regional and regional models indicates that a relatively simple model with two clusters and a spatially varying scale parameter can compete very well against complex models.

Investigating How Reproducibility and Geometrical Representation in UMAP Dimensionality Reduction Impact the Stratification of Breast Cancer Tumors

Advances in next-generation sequencing have provided high-dimensional RNA-seq datasets, allowing the stratification of some tumor patients based on their transcriptomic profiles. Machine learning methods have been used to reduce and cluster high-dimensional data. Recently, uniform manifold approximation and projection (UMAP) was applied to project genomic datasets in low-dimensional Euclidean latent space. Here, we evaluated how different representations of the UMAP embedding can impact the analysis of breast cancer (BC) stratification. We projected BC RNA-seq data on Euclidean, spherical, and hyperbolic spaces, and stratified BC patients via clustering algorithms. We also proposed a pipeline to yield more reproducible clustering outputs. The results show how the selection of the latent space can affect downstream stratification results and suggest that the exploration of different geometrical representations is recommended to explore data structure and samples’ relationships.

3D Tooth Instance Segmentation Learning Objectness and Affinity in Point Cloud

Digital dentistry has received more attention in the past decade. However, current deep learning-based methods still encounter difficult challenges. The proposal-based methods are sensitive to the localization results due to the lack of local cues, while the proposal-free methods have poor clustering outputs because of the affinity measured by the low-level characteristics, especially in situations of tightly arranged teeth. In this article, we present a novel proposal-based approach to combine objectness and pointwise knowledge in an attention mechanism for point cloud-based tooth instance segmentation, using local information to improve 3D proposal generation and measuring the importance of local points by calculating the center distance. We evaluate the performance of our approach by constructing a Shining3D tooth instance segmentation dataset. The experimental results verify that our approach gives competitive results when compared with the other available approaches.

The effect of renewable energy incorporation on power grid stability and resilience.

Contemporary proliferation of renewable power generation is causing an overhaul in the topology, composition, and dynamics of electrical grids. These low-output, intermittent generators are widely distributed throughout the grid, including at the household level. It is critical for the function of modern power infrastructure to understand how this increasingly distributed layout affects network stability and resilience. This paper uses dynamical models, household power consumption, and photovoltaic generation data to show how these characteristics vary with the level of distribution. It is shown that resilience exhibits daily oscillations as the grid’s effective structure and the power demand fluctuate. This can lead to a substantial decrease in grid resilience, explained by periods of highly clustered generator output. Moreover, the addition of batteries, while enabling consumer self-sufficiency, fails to ameliorate these problems. The methodology identifies a grid’s susceptibility to disruption resulting from its network structure and modes of operation.

Grouped SPME Comparison of Floral Scent as a Method of Unlocking Phylogenetic Patterns in Volatiles

Global crop production rate has exceeded the availability of pollination services provided by managed honeybees, and habitat loss remains a key factor in the loss of wild pollinators. Revegetation of agricultural land and wild pollination may provide a solution; however, the collection of floral trait data that are correlated to pollinator preferences remains an under studied and complex process. Here, we demonstrate a method for scent analysis, ordination [non-metric dimensional scaling (NMDS)], and clustering outputs that provides a fast and reproducible procedure for a broad grouping of flora based on scent and unlocking characteristic inter-floral patterns. We report the floral profiles of 15 unstudied native Australian plant species and the extent to which they match the commonly cultivated seed crops of Daucus carota L and Brassica rapa L. Through solid-phase microextraction (SPME) paired with gas chromatography–mass spectrometry (GC-MS), we identify a set of inter-family shared, common floral volatiles from these plant species as well as unique and characteristic patterns.

COMPARATIVE STUDY OF DISTANCE MEASURES ON FUZZY SUBTRACTIVE CLUSTERING

Clustering is a data analysis process which applied to classify the unlabeled data. Fuzzy clustering is a clustering method based on membership value which enclosing set of fuzzy as a measurement base for classification process. Fuzzy Subtractive Clustering (FSC) is included in one of fuzzy clustering method. This research applies Hamming distance and combined Minkowski Chebysev distance as a distance parameter in Fuzzy Subtractive Clustering. The objective of this research is to compare the output quality of the cluster from Fuzzy Subtractive Clustering by using Hamming distance and combine Minkowski Chebysev distance. The comparison of the two distances aims to see how well the clusters are produced from two different distances. The data used is data on hypertension. The variables used are age, gender, systolic pressure, diastolic pressure, and body weight. This research shows that the Partition Coefficient value resulted on Fuzzy Subtractive Clustering by applying combined Minkowski Chebysev distance is higher than the application of Hamming distance. Based on this, it can be concluded that in this study the quality of the cluster output using the combined Minkowski Chebysev distance is better.

Improved Density Peaking Algorithm for Community Detection Based on Graph Representation Learning

There is a large amount of information in the network data that we can exploit. It is difficult for classical community detection algorithms to handle network data with sparse topology. Representation learning of network data is usually paired with clustering algorithms to solve the community detection problem. Meanwhile, there is always an unpredictable distribution of class clusters output by graph representation learning. Therefore, we propose an improved density peak clustering algorithm (ILDPC) for the community detection problem, which improves the local density mechanism in the original algorithm and can better accommodate class clusters of different shapes. And we study the community detection in network data. The algorithm is paired with the benchmark model Graph sample and aggregate (GraphSAGE) to show the adaptability of ILDPC for community detection. The plotted decision diagram shows that the ILDPC algorithm is more discriminative in selecting density peak points compared to the original algorithm. Finally, the performance of K-means and other clustering algorithms on this benchmark model is compared, and the algorithm is proved to be more suitable for community detection in sparse networks with the benchmark model on the evaluation criterion F1-score. The sensitivity of the parameters of the ILDPC algorithm to the low-dimensional vector set output by the benchmark model GraphSAGE is also analyzed.

Cluster output synchronization for memristive neural networks

Herein, cluster output synchronization for memristive neural networks (MNNs) is investigated using two different control schemes. Existing synchronization models for MNNs focus on the behavior of a single neuron node in one-cluster networks. However, actual neural networks (NNs) are clustered organizations consisting of multiple interacting clusters, where the nodes from the same cluster combine and work together. This study proposes a cluster output synchronization model for MNNs, which considers the combination output behavior of the nodes in NNs clusters. Accordingly, two specific control schemes are designed: one based on feedback control involves designing a small number of controllers to reduce control costs, and the other based on adaptive control involves designing multiple adjustable controllers to increase the anti-interference capacity of the control system. Meanwhile, to facilitate synchronization in MNNs, a model relationship between MNNs and traditional NNs is investigated. By utilizing the control schemes, model relationship, and Lyapunov stability theory, sufficient conditions are obtained for validating the cluster output synchronization. Finally, several numerical examples are given to illustrate the accuracy of the theoretical results.