High-quality Clusters Research Articles

Bioinspired Hybrid and Incomplete Data Clustering

Enhancing the results of data clustering for hybrid (numeric and categorical) and missing data is paramount for robust pattern recognition and decision-making in various fields. Traditional clustering algorithms often need help with heterogeneous data types and incomplete information, leading to suboptimal groupings and potentially biased insights. By addressing these challenges, advanced techniques, such as bioinspired algorithms, can provide more accurate and comprehensive clustering. We show the usefulness of using internal validity indices in obtaining high-quality clusters for hybrid and incomplete data. In addition, this work analyzes the influence of applying the recently proposed PAntSA on the performance of hybrid and incomplete clustering algorithms. To do this, the results of different algorithms are compared before and after applying PAntSA. The statistical analysis of the results provides experimental evidence that supports that the PAntSA algorithm improves the quality of the groups obtained by hybrid and incomplete data clustering methods.

Staleness-Reduction Mini-Batch K-Means.

K -means (km) is a clustering algorithm that has been widely adopted due to its simple implementation and high clustering quality. However, the standard km suffers from high computational complexity and is therefore time-consuming. Accordingly, the mini-batch (mbatch) km is proposed to significantly reduce computational costs in a manner that updates centroids after performing distance computations on just a mbatch, rather than a full batch, of samples. Even though the mbatch km converges faster, it leads to a decrease in convergence quality because it introduces staleness during iterations. To this end, in this article, we propose the staleness-reduction mbatch (srmbatch) km, which achieves the best of two worlds: low computational costs like the mbatch km and high clustering quality like the standard km. Moreover, srmbatch still exposes massive parallelism to be efficiently implemented on multicore CPUs and many-core GPUs. The experimental results show that srmbatch can converge up to 40× - 130× faster than mbatch when reaching the same target loss, and srmbatch is able to reach 0.2%-1.7% lower final loss than that of mbatch.

Interpretable sequence clustering

Categorical sequence clustering is vital across various domains; however, the interpretability of cluster assignments presents considerable challenges. Sequences inherently lack explicit features, and existing sequence clustering algorithms heavily rely on complex representations, which complicates the explanation of their outcomes. To address this issue, we propose a method called Interpretable Sequence Clustering Tree (ISCT), which combines sequential patterns with a concise and interpretable tree structure. ISCT leverages k−1 patterns to generate k leaf nodes, corresponding to k clusters, which provides an intuitive explanation on how each cluster is formed. More precisely, ISCT first projects sequences into random subspaces and then utilizes the k-means algorithm to obtain high-quality initial cluster assignments. Subsequently, it constructs a pattern-based decision tree using a boosting strategy in which sequences are re-projected and re-clustered at each node before mining the top-1 discriminative splitting pattern. Experimental results on 14 real-world data sets demonstrate that our proposed method provides an interpretable tree structure while delivering fast and accurate cluster assignments.

Constrained Density-Based Spatial Clustering of Applications with Noise (DBSCAN) using hyperparameter optimization

This article proposes a hyperparameter optimization method for density-based spatial clustering of applications with noise (DBSCAN) with constraints, termed HC-DBSCAN. While DBSCAN is effective at creating non-convex clusters, it cannot limit the number of clusters. This limitation is difficult to address with simple adjustments or heuristic methods. We approach constrained DBSCAN as an optimization problem and solve it using a customized alternating direction method of multipliers Bayesian optimization (ADMMBO). Our custom ADMMBO enables HC-DBSCAN to reuse clustering results for enhanced computational efficiency, handle integer-valued parameters, and incorporate constraint functions that account for the degree of violations to improve clustering performance. Furthermore, we propose an evaluation metric, penalized Davies–Bouldin score, with a computational cost of O(N). This metric aims to mitigate the high computational cost associated with existing metrics and efficiently manage noise instances in DBSCAN. Numerical experiments demonstrate that HC-DBSCAN, equipped with the proposed metric, generates high-quality non-convex clusters and outperforms benchmark methods on both simulated and real datasets.

Density-based clustering with differential privacy

In recent years, differentially private clustering has received increasing attention. However, most existing differentially private clustering algorithms cannot achieve better results when handling non-convex datasets. To enhance knowledge extraction from data while protecting users' sensitive information, we propose a density-based clustering algorithm with differential privacy. Specifically, we incorporate differential privacy mechanisms into the density-based clustering paradigm to enhance the effectiveness of differentially private clustering on non-convex datasets. Firstly, to avoid privacy leakage, we employ the Laplace mechanism for inject noise into the density during the density estimation stage. Then, we design a privacy budget allocation scheme in the cluster expansion stage to make it harder for attackers to access private information. Theoretical analysis demonstrates that our algorithm satisfies ϵ-differential privacy. Experimental outcomes in synthetic and real-world datasets show that our introduced algorithm can obtain high-quality clustering results when dealing with non-convex datasets. In the approximation experiments, it is evident that our algorithm outperforms others in terms of approximation.

Federated Momentum Contrastive Clustering

Self-supervised representation learning and deep clustering are mutually beneficial to learn high-quality representations and cluster data simultaneously in centralized settings. However, it is not always feasible to gather large amounts of data at a central entity, considering data privacy requirements and computational resources. Federated Learning (FL) has been developed successfully to aggregate a global model while training on distributed local data, respecting the data privacy of edge devices. However, most FL research effort focuses on supervised learning algorithms. A fully unsupervised federated clustering scheme has not been considered in the existing literature. We present federated momentum contrastive clustering (FedMCC), a generic federated clustering framework that can not only cluster data automatically but also extract discriminative representations training from distributed local data over multiple users. In FedMCC, we demonstrate a two-stage federated learning paradigm where the first stage aims to learn differentiable instance embeddings and the second stage accounts for clustering data automatically. The experimental results show that FedMCC not only achieves superior clustering performance but also outperforms several existing federated self-supervised methods for linear evaluation and semi-supervised learning tasks. Additionally, FedMCC can easily be adapted to ordinary centralized clustering through what we call momentum contrastive clustering (MCC). We show that MCC achieves state-of-the-art clustering accuracy results in certain datasets such as STL-10 and ImageNet-10. We also present a method to reduce the memory footprint of our clustering schemes.

A New Hybrid Optimization Algorithm to Solve the Clustering Problem

: This paper presents a novel hybrid technique, merging the K-means algorithm with Genetic Algorithm (GA), aiming to enhance clustering performance. This approach leverages the strengths of both algorithms, enabling improved cluster generation by overcoming individual algorithmic limitations. The GA-KM algorithm is introduced to aid K-means in avoiding local optima, with GA exhibiting proficiency in determining optimal cluster initialization and parameter optimization. The focus is on developing a GA-based algorithm for generating high-quality clusters efficiently. Notably, the research explores the application of this hybrid approach to address issues in the educational domain, specifically for out-of-school children. The fitness function in GA is tailored to the problem area, emphasizing the need for an appropriate system to study and address school children's problems. The research proposes a hybrid algorithm (KM-GA-NM-PSO) that amalgamates the best features of existing algorithms, thereby overcoming individual limitations and promising superior results. This hybridization is expected to yield high-quality clusters with minimal function evaluations, outperforming other methods by producing clusters with small standard deviations on selected datasets. The proposed approach, combining KM, GA, NM, and PSO algorithms, demonstrates improved data clustering quality and algorithmic efficiency.

Predictive Maintenance for Electrical Substation Components Using K-Means Clustering: A Case Study

PT PLN (Persero) UP2D Kalselteng aims to provide reliable electricity supply, necessitating effective substation maintenance. This study proposes a predictive maintenance approach using K-means clustering on electrical current performance data from eight components in the Amuntai main electrical substation. The data undergoes preprocessing, including mapping to absolute z-scores to address electricity fluctuations. The K-means algorithm clusters performances, and models are evaluated using Silhouette scores. Results indicate the potential for predicting maintenance needs, as clusters align with real power outage data. The proposed method provides a proactive strategy for substation maintenance, enhancing system reliability. Feature combination experiments reveal that individual models for transformers and feeders are optimal. Hyperparameter tuning refines models, showcasing silhouette scores above 0.5, indicative of high-quality clusters. Comparisons with real-world power outage data validate the model's capability to identify anomalies, reinforcing the feasibility of the predictive maintenance approach. While the study demonstrates promise, on-field implementation and additional experiments are crucial for comprehensive validation and refinement of the predictive maintenance models.

A Non-parametric Graph Clustering Framework for Multi-View Data

Multi-view graph clustering (MVGC) derives encouraging grouping results by seamlessly integrating abundant information inside heterogeneous data, and has captured surging focus recently. Nevertheless, the majority of current MVGC works involve at least one hyper-parameter, which not only requires additional efforts for tuning, but also leads to a complicated solving procedure, largely harming the flexibility and scalability of corresponding algorithms. To this end, in the article we are devoted to getting rid of hyper-parameters, and devise a non-parametric graph clustering (NpGC) framework to more practically partition multi-view data. To be specific, we hold that hyper-parameters play a role in balancing error item and regularization item so as to form high-quality clustering representations. Therefore, under without the assistance of hyper-parameters, how to acquire high-quality representations becomes the key. Inspired by this, we adopt two types of anchors, view-related and view-unrelated, to concurrently mine exclusive characteristics and common characteristics among views. Then, all anchors' information is gathered together via a consensus bipartite graph. By such ways, NpGC extracts both complementary and consistent multi-view features, thereby obtaining superior clustering results. Also, linear complexities enable it to handle datasets with over 120000 samples. Numerous experiments reveal NpGC's strong points compared to lots of classical approaches.

PPPCT: Privacy-Preserving framework for Parallel Clustering Transcriptomics data

Single-cell transcriptomics data provides crucial insights into patients’ health, yet poses significant privacy concerns. Genomic data privacy attacks can have deep implications, encompassing not only the patients’ health information but also extending widely to compromise their families’. Moreover, the permanence of leaked data exacerbates the challenges, making retraction an impossibility. While extensive efforts have been directed towards clustering single-cell transcriptomics data, addressing critical challenges, especially in the realm of privacy, remains pivotal. This paper introduces an efficient, fast, privacy-preserving approach for clustering single-cell RNA-sequencing (scRNA-seq) datasets. The key contributions include ensuring data privacy, achieving high-quality clustering, accommodating the high dimensionality inherent in the datasets, and maintaining reasonable computation time for big-scale datasets. Our proposed approach utilizes the map-reduce scheme to parallelize clustering, addressing intensive calculation challenges. Intel Software Guard eXtension (SGX) processors are used to ensure the security of sensitive code and data during processing. Additionally, the approach incorporates a logarithm transformation as a preprocessing step, employs non-negative matrix factorization for dimensionality reduction, and utilizes parallel k-means for clustering. The approach fully leverages the computing capabilities of all processing resources within a secure private cloud environment. Experimental results demonstrate the efficacy of our approach in preserving patient privacy while surpassing state-of-the-art methods in both clustering quality and computation time. Our method consistently achieves a minimum of 7% higher Adjusted Rand Index (ARI) than existing approaches, contingent on dataset size. Additionally, due to parallel computations and dimensionality reduction, our approach exhibits efficiency, converging to very good results in less than 10 seconds for a scRNA-seq dataset with 5000 genes and 6000 cells when prioritizing privacy and under two seconds without privacy considerations.Availability and implementationCode and datasets availability: https://github.com/University-of-Windsor/PPPCT.

Augmented weighted K-means grey wolf optimizer: An enhanced metaheuristic algorithm for data clustering problems

This study presents the K-means clustering-based grey wolf optimizer, a new algorithm intended to improve the optimization capabilities of the conventional grey wolf optimizer in order to address the problem of data clustering. The process that groups similar items within a dataset into non-overlapping groups. Grey wolf hunting behaviour served as the model for grey wolf optimizer, however, it frequently lacks the exploration and exploitation capabilities that are essential for efficient data clustering. This work mainly focuses on enhancing the grey wolf optimizer using a new weight factor and the K-means algorithm concepts in order to increase variety and avoid premature convergence. Using a partitional clustering-inspired fitness function, the K-means clustering-based grey wolf optimizer was extensively evaluated on ten numerical functions and multiple real-world datasets with varying levels of complexity and dimensionality. The methodology is based on incorporating the K-means algorithm concept for the purpose of refining initial solutions and adding a weight factor to increase the diversity of solutions during the optimization phase. The results show that the K-means clustering-based grey wolf optimizer performs much better than the standard grey wolf optimizer in discovering optimal clustering solutions, indicating a higher capacity for effective exploration and exploitation of the solution space. The study found that the K-means clustering-based grey wolf optimizer was able to produce high-quality cluster centres in fewer iterations, demonstrating its efficacy and efficiency on various datasets. Finally, the study demonstrates the robustness and dependability of the K-means clustering-based grey wolf optimizer in resolving data clustering issues, which represents a significant advancement over conventional techniques. In addition to addressing the shortcomings of the initial algorithm, the incorporation of K-means and the innovative weight factor into the grey wolf optimizer establishes a new standard for further study in metaheuristic clustering algorithms. The performance of the K-means clustering-based grey wolf optimizer is around 34% better than the original grey wolf optimizer algorithm for both numerical test problems and data clustering problems.

A Simultaneous Wireless Information and Power Transfer-Based Multi-Hop Uneven Clustering Routing Protocol for EH-Cognitive Radio Sensor Networks

Clustering protocols and simultaneous wireless information and power transfer (SWIPT) technology can solve the issue of imbalanced energy consumption among nodes in energy harvesting-cognitive radio sensor networks (EH-CRSNs). However, dynamic energy changes caused by EH/SWIPT and dynamic spectrum availability prevent existing clustering routing protocols from fully leveraging the advantages of EH and SWIPT. Therefore, a multi-hop uneven clustering routing protocol is proposed for EH-CRSNs utilizing SWIPT technology in this paper. Specifically, an EH-based energy state function is proposed to accurately track the dynamic energy variations in nodes. Utilizing this function, dynamic spectrum availability, neighbor count, and other information are integrated to design the criteria for selecting high-quality cluster heads (CHs) and relays, thereby facilitating effective data transfer to the sink. Intra-cluster and inter-cluster SWIPT mechanisms are incorporated to allow for the immediate energy replenishment for CHs or relays with insufficient energy while transmitting data, thereby preventing data transmission failures due to energy depletion. An energy status control mechanism is introduced to avoid the energy waste caused by excessive activation of the SWIPT mechanism. Simulation results indicate that the proposed protocol markedly improves the balance of energy consumption among nodes and enhances network surveillance capabilities when compared to existing clustering routing protocols.

Deciphering spatial domains from spatially resolved transcriptomics with Siamese graph autoencoder.

Cell clustering is a pivotal aspect of spatial transcriptomics (ST) data analysis as it forms the foundation for subsequent data mining. Recent advances in spatial domain identification have leveraged graph neural network (GNN) approaches in conjunction with spatial transcriptomics data. However, such GNN-based methods suffer from representation collapse, wherein all spatial spots are projected onto a singular representation. Consequently, the discriminative capability of individual representation feature is limited, leading to suboptimal clustering performance. To address this issue, we proposed SGAE, a novel framework for spatial domain identification, incorporating the power of the Siamese graph autoencoder. SGAE mitigates the information correlation at both sample and feature levels, thus improving the representation discrimination. We adapted this framework to ST analysis by constructing a graph based on both gene expression and spatial information. SGAE outperformed alternative methods by its effectiveness in capturing spatial patterns and generating high-quality clusters, as evaluated by the Adjusted Rand Index, Normalized Mutual Information, and Fowlkes-Mallows Index. Moreover, the clustering results derived from SGAE can be further utilized in the identification of 3-dimensional (3D) Drosophila embryonic structure with enhanced accuracy. Benchmarking results from various ST datasets generated by diverse platforms demonstrate compelling evidence for the effectiveness of SGAE against other ST clustering methods. Specifically, SGAE exhibits potential for extension and application on multislice 3D reconstruction and tissue structure investigation. The source code and a collection of spatial clustering results can be accessed at https://github.com/STOmics/SGAE/.

Research on e-commerce user segmentation and customized marketing strategy based on cluster analysis

Abstract E-commerce user segmentation is the basis for enterprises to accurately formulate marketing strategies and successfully manage their customer base. With the rapid development of e-commerce, this paper improves the traditional K-means clustering algorithm and proposes a SAPK-means algorithm, which effectively excludes the noisy data and isolated points in the dataset and obtains the high-quality initial clustering center. Company A’s e-commerce platform is used to apply the SAPK-means algorithm for customer segmentation, and the results are analyzed in detail before proposing targeted marketing strategies. The customized marketing strategy’s effect is evaluated through sales and platform user satisfaction. Through experimental testing, this paper concludes that the five types of segmented customer groups account for 9.46%, 18.43%, 36.95%, 21.91%, and 13.24% of the total number of samples, respectively, known as the “platinum customer group”, “gold customer group “Platinum customer group”, “Gold customer group”, “Silver customer group”, “Copper customer group” and “Iron customer group”, respectively.

A Deep Diagnostic Framework Using Explainable Artificial Intelligence and Clustering.

An important part of diagnostics is to gain insight into properties that characterize a disease. Machine learning has been used for this purpose, for instance, to identify biomarkers in genomics. However, when patient data are presented as images, identifying properties that characterize a disease becomes far more challenging. A common strategy involves extracting features from the images and analyzing their occurrence in healthy versus pathological images. A limitation of this approach is that the ability to gain new insights into the disease from the data is constrained by the information in the extracted features. Typically, these features are manually extracted by humans, which further limits the potential for new insights. To overcome these limitations, in this paper, we propose a novel framework that provides insights into diseases without relying on handcrafted features or human intervention. Our framework is based on deep learning (DL), explainable artificial intelligence (XAI), and clustering. DL is employed to learn deep patterns, enabling efficient differentiation between healthy and pathological images. Explainable artificial intelligence (XAI) visualizes these patterns, and a novel "explanation-weighted" clustering technique is introduced to gain an overview of these patterns across multiple patients. We applied the method to images from the gastrointestinal tract. In addition to real healthy images and real images of polyps, some of the images had synthetic shapes added to represent other types of pathologies than polyps. The results show that our proposed method was capable of organizing the images based on the reasons they were diagnosed as pathological, achieving high cluster quality and a rand index close to or equal to one.

Clustering of diet, physical activity and sedentary behaviour and related physical and mental health outcomes: a systematic review

BackgroundPhysical activity (PA), sedentary behaviour (SB) and diet play an important role in the physical and mental health of young people. Understanding how these behaviours cluster, and the impact of clusters on health is important for the development of public health interventions. This review examines the prevalence of clusters of PA, sedentary time, and dietary behaviours, and how clusters relate to physical and mental health indicators among children, adolescents and young adults.MethodsElectronic (PubMed, Web of Science and Scopus) and manual searches were conducted for articles that were (i) observational studies including children, adolescents and/or young adults aged 5–24 years, (ii) examined the 'patterning', ‘clustering’, or ‘co-existence’ of each of PA, dietary behaviour and SB, and (iii) published in English up to and including July 2022. In addition to information on clustering, data on physical and mental health outcomes were extracted where reported. Included studies were assessed using the Cochrane risk of bias for observational studies. A narrative synthesis was conducted due to high heterogeneity. This review was registered with PROSPERO (CRD42021230976).ResultsForty-nine cross-sectional studies and four prospective cohort studies from eighteen countries reporting data from 778,415 individual participants were included. A broad range of clusters (n = 172) were found (healthy, unhealthy, and mixed). Mixed clusters were common (n = 98), and clusters of high diet quality, low PA and high SB were more prevalent in girls, while mixed clusters of high PA, high SB and low diet quality were more prevalent in boys. Unhealthy clusters comprising low moderate to vigorous PA, low consumption of fruits and vegetables, and high screen time were prevalent, particularly in those from lower socioeconomic status families. Compared to those with healthy behavioural clusters, those with unhealthy and mixed clusters had a higher adiposity, higher risk of cardiovascular disease, poorer mental health scores, and lower cardiorespiratory fitness.ConclusionsPA, SB and diet cluster in healthy, unhealthy and mixed patterns in young people that differ across sociodemographic characteristics. Unhealthy clusters are associated with poorer health outcomes. Intervention strategies targeting un-clustering multiple unhealthy behaviours should be developed and evaluated for their impact on health outcomes.

A Convolutional Autoencoder-based Explainable Clustering Approach for Resting-State EEG Analysis.

Machine learning methods have frequently been applied to electroencephalography (EEG) data. However, while supervised EEG classification is well-developed, relatively few studies have clustered EEG, which is problematic given the potential for clustering EEG to identify novel subtypes or patterns of dynamics that could improve our understanding of neuropsychiatric disorders. There are established methods for clustering EEG using manually extracted features that reduce the richness of the feature space for clustering, but only a couple studies have sought to use deep learning-based approaches with automated feature learning to cluster EEG. Those studies involve separately training an autoencoder and then performing clustering on the extracted features, and the separation of those steps can lead to poor quality clustering. In this study, we propose an explainable convolutional autoencoder-based approach that combines model training with clustering to yield high quality clusters. We apply the approach within the context of schizophrenia (SZ), identifying 8 EEG states characterized by varying levels of δ activity. We also find that individuals who spend more time outside of the dominant state tend to have increased negative symptom severity. Our approach represents a significant step forward for clustering resting-state EEG data and has the potential to lead to novel findings across a variety of neurological and neuropsychological disorders in future years.

Reducing Uncertainty and Increasing Confidence in Unsupervised Learning

This paper presents the development of a novel algorithm for unsupervised learning called RUN-ICON (Reduce UNcertainty and Increase CONfidence). The primary objective of the algorithm is to enhance the reliability and confidence of unsupervised clustering. RUN-ICON leverages the K-means++ method to identify the most frequently occurring dominant centres through multiple repetitions. It distinguishes itself from existing K-means variants by introducing novel metrics, such as the Clustering Dominance Index and Uncertainty, instead of relying solely on the Sum of Squared Errors, for identifying the most dominant clusters. The algorithm exhibits notable characteristics such as robustness, high-quality clustering, automation, and flexibility. Extensive testing on diverse data sets with varying characteristics demonstrates its capability to determine the optimal number of clusters under different scenarios. The algorithm will soon be deployed in real-world scenarios, where it will undergo rigorous testing against data sets based on measurements and simulations, further proving its effectiveness.

Gaia colour-magnitude diagrams of young open clusters

Context. Automated analyses of Gaia astrometric data have led to the discovery of many new high-quality open cluster candidates. When a good determination of their parameters is available, these objects become excellent tools for investigating the properties of our Galaxy. Aims. We explore whether young open clusters can be readily identified from Gaia data alone by studying the properties of their Gaia colour-magnitude diagrams. We also want to compare the results of a traditional cluster analysis with those of automated methods. Methods. We selected three young open cluster candidates from the UBC catalogue, ranging from a well-populated object with a well-defined sequence to a poorly populated and poorly defined candidate. We obtained classification spectra for the brightest stars in each. We redetermined members based on EDR3 data and fitted isochrones to derive the age, distance, and reddening. Results. All three candidates are real clusters with ages below 100 Ma. UBC 103 is a moderately populous cluster, with an age around 70 Ma. At a distance of ∼3 kpc, it forms a binary cluster with nearby NGC 6683. UBC 114 is a relatively proximal (∼1.5 kpc) poorly populated cluster containing two early-B stars. UBC 587 is a dispersed and very young (≤10 Ma) cluster located at ∼3 kpc, behind the Cygnus X region, which may be a valuable tracer of the Orion arm. Conclusions. The OCfinder methodology for the identification of new open clusters is extremely effective, with even poor candidates resulting in interesting detections. The presence of an almost vertical photometric sequence in the Gaia colour-magnitude diagram is a safe way to identify young open clusters. Automated methods for the determination of cluster properties give approximate solutions, but these are still subject to difficulties. There is some evidence suggesting that artificial intelligence systems may systematically underestimate extinction, which may impact in the age determination.

Research on the Construction of Entrepreneurship Course Clusters for Students in Agriculture and Forestry Universities

Curriculum construction is the key factor in determining the quality of education. The innovation and entrepreneurship course clusters have effectively upgraded entrepreneurship courses in terms of talent cultivation, resource integration, and educational practice. In response to the existing problems and difficulties in the construction of entrepreneurship course clusters in agricultural and forestry universities, suggestions and measures are proposed from macro-, meso-, and micro-level perspectives to reach a consensus on entrepreneurship education, construct interdisciplinary course clusters, and improve the quality and effect of education implementation. This provides a reference for agricultural and forestry universities to construct high-quality entrepreneurship course clusters.