Optimal Number Of Clusters Research Articles

히스토그램 자료를 위한 성긴 k-평균 군집분석에 관한 연구

In this paper, we investigate a sparse k-means clustering method for histogram-valued data. The distances between histogram-valued observations are defined using the Wasserstein-Kantorovich distances to group p-dimensional histogram-valued data. Clustering is performed using the sparse k-means clustering method with the distance matrix computed for each dimension. The proposed method maximizes the weighted sums of squared distances between clusters. For various value of k, we apply the sparse k-means clustering method and determine the optimal number of clusters with the Silhouette measure. Simulation studies were conducted to compare the proposed method with the standard k-means clustering method in terms of cluster agreement and selected variables. Additionally, we analyzed real data from the monthly average temperatures of 48 US states. As a result of numerical analysis, it was confirmed that the proposed method shows superior performance and effective variable selection.

Multidimensional Dynamic Destination Recommender Search System Employing Clustering: A Machine Learning Approach

Objectives: Recommender Systems (RS) powered by algorithms of machine learning is a popular tool for planning and implementing custom-made travel proficiencies. The persistence of this study is to recommend destinations according to a selection of various dimensions by the user. Methods: This approach uses a hybrid filtering system for recommendation with a weighted K-means clustering algorithm. For this study dataset was taken from Kaggle. Data considers different cities of India with different dimensions like city, name, type, and significance. According to the city first find latitude and longitude for precise clustering. Future work will incorporate optimization techniques to improve cluster formation recommendation accuracy. Findings: Clustering (unsupervised learning) is a separation technique that involves assigning locations to corresponding subsets of related clusters. The weighted K-means clustering algorithm is used with the elbow method which is used for discovering the optimum number of clusters. In weighted K-means algorithm for clustering uses scaling factor wi ​which transforms the impression of individual features to the whole distance calculation. It signifies the meaning of the ith feature in the perspective of the grouping task. Offering a scaling factor permits additional tractability in modifying the outcome of specific features on the distance calculation. It enables customization of the distance metric constructed on the specific requirements and characteristics of the records and clustering task. In this study, user can select multiple dimensions of their choice and get recommendations according to their choice. The proposed weighted K-means algorithm shows a significant improvement in accuracy which considers the proportion of correct recommendations out of all recommendations. A comparison with traditional K-means was conducted, where the weighted algorithm achieved a 17% higher accuracy due to its ability to give importance to specific features. The future version of the proposed system will incorporate optimization techniques for enhanced performance. Novelty: The suggested solution in this paper demonstrates that the user can enter the city of their choice. The recommended method indicates the city and nearby predilections once the user has selected their parameters, such as consuming formations or name or type. The ratio of relevant destinations that have been successfully recommended is 18% more compared to the K-means clustering algorithm. Keywords: Recommender System, Clustering, Destination Recommender System (DRS), Machine Learning, Weighted K-means clustering Algorithm

Time-series clustering of raw recordings of 24-hour ambulatory blood presssure for risk stratification in general population

Abstract Objective 24-hour ambulatory blood pressure measurements (ABPM) is the best out-of-office BP measurement used to refine cardiovascular (CV) risk stratification. Currently, only summary ABPM indexes are used in the clinical practice while the temporal characteristics of the raw BP recordings have not been explored. Therefore, in this study we employed an unsupervised machine learning (ML) algorithm suitable for multivariate time-series analysis to identify distinct BP and heart rate (HR) profiles associated with the incidence of adverse CV events. Design and Method In 1391 community-dwelling individuals (mean age, 46.9 years; 50.97% women), we acquired 24-hour ABPM and clinical data and collected CV events (n=399) on average 22 years later. We employed multivariate dynamic time warping and k-medoids algorithm on the raw recordings of systolic and diastolic BP and HR. To assess the clinical significance of the derived clusters, we compared the clinical characteristics and the incidence of adverse events. We validated the trained model using 24-hour ABPM obtained from external population cohort (n=1137). Results The silhouette score and Dunn index showed that the optimal number of clusters was 4. Across all clusters we observed significant difference in age and sex distribution. Cluster 4 had the worst CV risk factors profiling, with higher office BP and anti-hypertensive medication intake compared to the other clusters. We observed differences in the incidence of adverse events between clusters (Figure), with cluster 1 representing the lowest risk group and cluster 4 the highest. After adjusting for traditional CV risk factors, including office systolic BP, cluster 1 had the lowest risk (HR:0.92; 95%CI:0.84-0.99; P=0.036) and cluster 4 the highest (HR:1.13; 95%CI:1.03-1.24; P=0.006) as compared to the average population risk. The analysis of the external validation cohort revealed similar ABPM patterns and clinical characteristics. Conclusion Employing an unsupervised ML model on the raw ABPM recordings we identified clinically meaningful clusters associated with an increasing cumulative incidence of CV events. This ML analysis paves the way towards utilization of temporal BP and HR information and subsequently optimization of CV risk stratification.

Rationalised experiment design for parameter estimation with sensitivity clustering

Quantitative experiments are essential for investigating, uncovering, and confirming our understanding of complex systems, necessitating the use of effective and robust experimental designs. Despite generally outperforming other approaches, the broader adoption of model-based design of experiments (MBDoE) has been hindered by oversimplified assumptions and computational overhead. To address this, we present PARameter SEnsitivity Clustering (PARSEC), an MBDoE framework that identifies informative measurable combinations through parameter sensitivity (PS) clustering. We combined PARSEC with a new variant of Approximate Bayesian Computation-based parameter estimation for rapid, automated assessment and ranking of experiment designs. Using two kinetic model systems with distinct dynamical features, we show that PARSEC-based experiments improve the parameter estimation of a complex system. By its inherent formulation, PARSEC can account for experimental restrictions and parameter variability. Moreover, we demonstrate that there is a strong correlation between sample size and the optimal number of PS clusters in PARSEC, offering a novel method to determine the ideal sampling for experiments. This validates our argument for employing parameter sensitivity in experiment design and illustrates the potential to leverage both model architecture and system dynamics to effectively explore the experimental design space.

Multivariate time-series clustering of cardiopulmonary exercise test data for cardiovascular risk stratification

Abstract Aims Current clinical practice assesses cardiorespiratory fitness (CRF) by evaluating summary indexes of cardiopulmonary exercise tests (CPET). Yet, the raw time series recordings may hold additional information with clinical relevance. In this study, we investigated if the interpretation of raw CPET data could identify distinct CRF phenogroups and improve risk stratification. Methods In 1399 participants (mean age, 56.4 years; 37.7% women), who underwent maximal CPET by cycle ergometer, we collected baseline clinical characteristics and information on incident cardiovascular (CV) events (n=297) on average 4.3 years later. We employed dynamic time warping combined with k-medoids to identify phenogroups from raw CPET time-series. To evaluate the clinical significance of the derived phenogroups we compared clinical characteristics and incidence of CV events, while an external population cohort (n=171) was used to further validate the trained model. Results The optimal number of clusters was 5. We observed significant differences in age, use of medication, spirometry indexes and disease prevalence across all clusters. Cluster 5 was associated with the worst CV profile with higher use of antihypertensive medication and history of CV disease while cluster 1 represented the most favourable risk profile. After adjusting for traditional clinical covariables, clusters 4 (HR: 1.30; 95%CI: 1.02-1.65; p=0.033) and 5 (HR: 1.36; 95%CI: 1.08-1.72; p=0.0088) had significantly higher risk for incident CV events compared to clusters 1 and 2. Comparing clinical characteristics across clusters in the external validation cohort revealed similar patterns. Conclusion – Employing unsupervised machine learning on time series CPET recordings, we identified clinically meaningful CRF phenogroups. An integrative CRF profiling might facilitate CV risk stratification and consequently risk management.

Cluster-specific risk factors for suboptimal thromboprophylaxis in contemporary European cohort of patients with atrial fibrillation: the EHRA-PATHS Project

Abstract Background Oral anticoagulation (OAC) is essential to optimal thromboprophylaxis in patients with atrial fibrillation (AF). Owing to bleeding risk concerns, OAC use may be challenging in clinically complex high-risk AF patients. Comorbidities might lead to suboptimal OAC use. We assessed cluster-specific risk factors for non-prescribing OAC in pre-defined AF clusters in contemporary European EHRA-PATHS AF cohort. Methods The EHRA-PATHS AF cohort was constructed using data from the three prospective EurObservational Research Programme (EORP) AF registries - General Pilot, General Long-term, and AF III registry. A two-step cluster analysis was performed. A hierarchical cluster analysis with 18 variables was used to identify the optimal number of clusters, followed by a K-means cluster analysis to determine variable cluster association. Cluster-specific risk factors for non-prescribing OAC were identified using univariate and multivariate logistic regression analyses. Results Of 18.799 patients, 2086 (11.1%) were not prescribed with OAC. Among the three clusters (see Table), OAC non-prescribing was significantly more prevalent in Cluster 1 (younger healthier AF patients) compared with Cluster 2 (Elderly with less comorbidity, mostly non-cardiovascular) or Cluster 3 (Elderly with more comorbidity, mostly cardiovascular), both P<0.001, with no significant difference between Clusters 2 and 3 (P=0.161). Cluster-specific multivariable risk factors for non-prescribing OAC are shown in Table. In all three clusters AF clinical type was significantly associated with OAC non-use. In Cluster 1, OAC non-use was less likely in the presence of CHA2DS2-VASc stroke risk factors, while bleeding risk factors were predominant multivariable predictors of OAC non-use among elderly (Clusters 2 and 3). Conclusion Our findings in a large contemporary European cohort suggest that different considerations prevail in treatment decision on OAC use in AF patients in clinical practice, depending on patient’s age and AF clinical type, as well as the presence and type of underlying comorbidities. Our findings could help in formulating targeted interventions to optimize stroke prevention in different subgroups of AF patients.

Clustering Portuguese municipalities based on population health indicators

Abstract Background The large number of publicly available health-related indicators poses challenges in health planning to policymakers. Identifying patterns among municipalities may assist these processes, by grouping similar contexts and pointing out relevant differences among neighbouring locations. The aim of this work was to define clusters of municipalities based on health-related indicators. Methods All health-related indicators available at municipality level for the year 2021 were obtained from the Portuguese National Statistics Office website. Missing values were imputed with the closest available year or excluded. Indicators were standardized and K-means clustering was applied. Elbow criterion was used to select the optimal number of clusters. Sensitivity analysis was performed with a priori exclusion of indicators less directly associated with health outcomes. Results The clustering analysis revealed five distinct clusters of municipalities. Clusters were associated with specific socio-demographic patterns, providing insights into common health challenges and strengths within each group. Interestingly, geographic profiles emerged (e.g. coastal vs inland), despite no geographic information being specifically provided. One outlier municipality was identified (Lisbon), highlighting a need for special consideration. Conclusions Clustering techniques provide a data-driven approach to inform public health planning. Our findings show previously unrecognized patterns and relationships between municipalities. Using these clusters, policymakers may implement interventions based on what has been proven to work in similar contexts, enhance collaboration opportunities and promote a tailored policy development, contributing to an overall improvement of public health outcomes. Key messages • Clustering machine learning techniques provide a data-driven approach to inform public health planning. • Identifying comparable municipalities based on health-related indicators allows targeted public health interventions.

Phenotype groups in atrial cardiomyopathy and their prognosis

Abstract Background Atrial cardiomyopathy (ACM) includes patients with heterogenous demographic characteristics and comorbidities. Defining phenotype groups with similar clinical characteristics is important for the effective treatment of these patients, and for their prognosis and prevention of ACM. Purpose To conduct a cluster analysis that could help us in phenotyping ACM and exploring the clinical complexity of this entity. Methods We performed a hierarchical cluster analysis based on Ward’s Method using 10 clinical binary variables (hypertension, diabetes mellitus, heart failure, obesity, chronic obstructive pulmonary disease, cancer, coronary artery disease, sinus node dysfunction or atrio-ventricular block, implanted pacemaker, and ischemic stroke), identifying the optimal number of clusters. We investigated differences in demographic, laboratory and echocardiographic parameters, and treatment between defined clusters, and assessed the impact of these factors on prognosis. Results Of 724 evaluated consecutive patients with dilated left atrium only 200 met the criteria for advanced atrial cardiomyopathy, defined as severely dilated left atrium with volume index (LAVI) ≥ 48 ml/m2, preserved left ventricular systolic function - ejection fraction ≥ 50%, without an overt valvular or ventricular disease, and were included in the analysis. Mean age of the studied population was 73.91 ± 9.74 years, 58 % of them were women. We identified 4 clusters: cluster 1 (n = 73, 36%) were younger overweight patients with paroxysmal atrial fibrillation (AF); cluster 2 (n = 56, 28%) - old patients with congestive heart failure (CHF) and low body mass index (BMI); cluster 3 (n = 34, 17%)- diabetic patients with obesity and CHF; and cluster 4 (n = 37, 19%) old patients with tachycardia-bradycardia syndrome, implanted pacemakers, and long -standing AF. Over a mean follow up of 20.6 months cluster 2 had the highest mortality rate (28.6%), followed by cluster 3 (20.6%), compared to the lower mortality rate of cluster 1 and 4 (11% and 10.8%, respectively, p=0.047). The presence of heart failure (HR 4.4, CI 1.5 ÷12.7, p=0.006), cancer (HR 3.3, CI 1.6 ÷ 6.9, p=0.002) and greater than moderate tricuspid regurgitation (HR 5.4, CI 2.6 ÷ 11.3, p<0.001) were predictors of poor outcome. Conclusion In patients with advanced atrial cardiomyopathy, four main clusters were identified, differentiated by distinct comorbidities. Both clinical and echocardiographic parameters were found to be associated with an increased risk of mortality.Phenotype clusters of ACMKaplan-Meier curves of clusters

Cluster-specific predictors of early rhythm control in contemporary European cohort of patients with atrial fibrillation: The EHRA-PATHS Project

Abstract Background Early rhythm control (ERC) using antiarrhytmic drugs, electrical cardioversion or catheter ablation for AF within a year from first diagnosis of atrial fibrillation (AF) has been shown to reduce the risk of major adverse outcomes compared to usual care, and the effects of ERC have been suggested to be consistent across all AF subgroups with different risk profiles. We aimed to identify cluster-specific predictors of ERC in pre-defined AF clusters in contemporary European EHRA-PATHS AF cohort. Methods The EHRA-PATHS AF cohort was derived from the three prospective EurObservational Research Programme (EORP) AF registries - General Pilot, General Long-term, and AF III registry. A two-step cluster analysis was performed. A hierarchical cluster analysis with 18 variables was used to identify the optimal number of clusters, followed by a K-means cluster analysis to determine variable cluster association. Cluster-specific predictors of ERC were identified using univariate and multivariate logistic regression analyses. Results Of 18,799 patients, AF was diagnosed within 1 year in 8088 (43.0%), and 3913/8088 patients (48.4%) were treated with ERC. Among the three clusters - Cluster 1 (younger healthier AF patients) compared with Cluster 2 (Elderly with less comorbidity, mostly non-cardiovascular) and Cluster 3 (Elderly with more comorbidity, mostly cardiovascular), see Table, ERC was significantly more often used in Cluster 1 compared with Cluster 2 and Cluster 3 (both P<0.001), and in Cluster 2 compared with Cluster 3. Cluster-specific multivariable predictors of ERC are shown in Table. In all three clusters, asymptomatic AF (EHRA score I) and Northern Europe residence were inversely associated with ERC, while Western Europe residence was positively associated with ERC in Cluster 2 and 3. In Cluster 1, low stroke risk and hypertension were positively associated with ERC, while dilated left atrium (LA) and reduced left ventricular ejection fraction (LV EF) were negative predictors of ERC. In Cluster 2, age ≥80 years, enlarged LA, and CHA2DS2-VASc ≥2 for males and ≥3 for females were negatively associated with ERC. In Cluster 3, age ≥75 years, cardiomyopathy, liver disease, dementia and reduced LV EF were negative predictors of ERC. Conclusion Our study of a large European AF cohort showed that likelihood for ERC was greater in younger and healthier AF patients, as well as in elderly from Western Europe, while ERC was significantly less likely with increasing age and comorbidity (especially cardiovascular diseases). These findings likely reflect the lack of high-quality evidence supporting aggressive ERC in high-risk patients with AF. With increasing body of evidence favouring ERC such patients, a shift towards less selective consideration of ERC in daily practice may be expected.

Outcome and life expectancy associated with novel subgroups of atrial fibrillation: a data-driven cluster analysis

Abstract Background/Introduction Atrial fibrillation (AF) presents a significant global health burden, affecting millions and associated with elevated risks of mortality, heart failure, and stroke despite current treatment strategies. The current diagnosis of AF relies on episode characteristics, yet its heterogeneity calls for personalized assessment incorporating clinical features, biomarkers, and substrate determination for improved classification and management. Purpose This study aimed to develop a refined classification of AF according to comprehensive risk factors to identify patients at high-risk for poor prognosis and their life expectancy, aiming to facilitate personalized interventions. Methods A total of 7,391 participants aged 40-69 with AF at baseline, from UK Biobank, were recruited in the main analyses. Firstly, LASSO regression with Cox model and factor analysis were adopted for identification and integration of principal risk factors. Second, Consensus clustering analysis was employed to estimate the optimal cluster number and group all 7,391 participants. Difference in the risk of death and major complications, as well as reductions in life expectancy, among clusters were estimated. Replication was done in 2,399 participants with newly diagnosed AF within two years after baseline. Results Risk factors for AF were categorized into seven aspects (metabolic factors, respiratory factors, cardiovascular factors, renal and immunity diseases, mental health, acute illness, age), based on the correlation between them. Five distinct AF clusters were identified according to these seven characteristics: acute illness-related (Cluster 1), mental health-related (Cluster 2), cardiovascular disease-related (Cluster 3), immune-and-renal disease-related (Cluster 4), and respiratory-and-metabolism disease-related AF (Cluster 5). Patients with respiratory-and-metabolism disease-related AF had the highest risk of death, acute myocardial infarction, heart failure, and cerebral ischemic stroke, while those with acute illness-related AF had the lowest corresponding risk. In addition, compared with individuals with acute illness-related AF, those with respiratory-and-metabolism disease-related and mental health-related AF had the top 2 greatest loss of life expectancy. Furthermore, genetic variants for AF had different effect among the five clusters. Replication analysis confirmed the result stability. Conclusion A novel AF classification was developed, which provided insights into varying life expectancy and risks of death and complications among AF subgroups with distinct characteristics. It offers a practical approach for identifying high-risk patients, which may help to tailor precise interventions for AF management.Graphic abstractCluster features

Comprehensive Analysis of Phenotypic Traits in Chinese Cabbage Using 3D Point Cloud Technology

Studies on the phenotypic traits and their associations in Chinese cabbage lack precise and objective digital evaluation metrics. Traditional assessment methods often rely on subjective evaluations and experience, compromising accuracy and reliability. This study develops an innovative, comprehensive trait evaluation method based on 3D point cloud technology, with the aim of enhancing the precision, reliability, and standardization of the comprehensive phenotypic traits of Chinese cabbage. By using multi-view image sequences and structure-from-motion algorithms, 3D point clouds of 50 plants from each of the 17 Chinese cabbage varieties were reconstructed. Color-based region growing and 3D convex hull techniques were employed to measure 30 agronomic traits. Comparisons between 3D point cloud-based measurements of the plant spread, plant height, leaf area, and leaf ball volume and traditional methods yielded R2 values greater than 0.97, with root mean square errors of 1.27 cm, 1.16 cm, 839.77 cm3, and 59.15 cm2, respectively. Based on the plant spread and plant height, a linear regression prediction of Chinese cabbage weights was conducted, yielding an R2 value of 0.76. Integrated optimization algorithms were used to test the parameters, reducing the measurement time from 55 min when using traditional methods to 3.2 min. Furthermore, in-depth analyses including variation, correlation, principal component analysis, and clustering analyses were conducted. Variation analysis revealed significant trait variability, with correlation analysis indicating 21 pairs of traits with highly significant positive correlations and 2 pairs with highly significant negative correlations. The top six principal components accounted for 90% of the total variance. Using the elbow method, k-means clustering determined that the optimal number of clusters was four, thus classifying the 17 cabbage varieties into four distinct groups. This study provides new theoretical and methodological insights for exploring phenotypic trait associations in Chinese cabbage and facilitates the breeding and identification of high-quality varieties. Compared with traditional methods, this system provides significant advantages in terms of accuracy, speed, and comprehensiveness, with its low cost and ease of use making it an ideal replacement for manual methods, being particularly suited for large-scale monitoring and high-throughput phenotyping.

Data-driven Machine Learning Models for Risk Stratification and Prediction of Emergence Delirium in Pediatric Patients Underwent Tonsillectomy/Adenotonsillectomy.

In the pediatric surgical population, Emergence Delirium (ED) poses a significant challenge. This study aims to develop and validate machine learning (ML) models to identify key features associated with ED and predict its occurrence in children undergoing tonsillectomy or adenotonsillectomy. The analysis involved data cleaning, exploratory data analysis (EDA), supervised predictive modeling, and unsupervised learning on a medical dataset (n = 423). After preliminary data cleaning, EDA encompassed plotting histograms, boxplots, pairplots, and correlation heatmaps to understand variable distributions and relationships. Four predictive models were trained including logistic regression (LR), random forest (RF), Support Vector Machine (SVM), and Gradient Boosting (XGBoost). The models were evaluated and compared using Receiver Operating Characteristic (ROC) Area Under the Curve (AUC), precision, recall, and feature importance. The RF model showed better performance and was used for the test (AUC-ROC 0.96, precision 1.00, and recall 0.92 on the validation set). K-means clustering was applied to find groups within the data. Elbow method and silhouette scores were used to determine the optimal number of clusters. The formed clusters were analyzed by aggregating features to understand the characteristics of each cluster. EDA revealed significant positive correlations between age, weight, American Society of Anesthesiologists (ASA) health score, and surgery duration with the risk of developing ED. Among the ML models, RF achieved the highest performance. Key predictive variables, based on the model's feature importance, included delirium screening scales, extubation time, and time to regain consciousness. Unsupervised K-means clustering identified 2-3 optimal clusters, which represented distinct patient subgroups: younger, healthier, low-risk individuals (cluster 0), and older patients with increasing chronic disease burden, higher delirium screening scores, and consequently higher post-operative delirium risk (clusters 1 and 2). ML techniques are valuable tools for extracting insights and making accurate predictions from healthcare data. High-performing algorithm-based models can be implemented for clinical decision support systems, facilitating early identification and intervention for ED in pediatric patients. By investigating various variables, it is possible to assess risk and implement preventive measures effectively. Furthermore, unsupervised clustering reveals distinct patient subgroups, enabling personalized perioperative management strategies and enhancing overall patient care.

A Method for Predicting the Impact of Labor Education at Institutions of Higher Education Based on a Two-Stage Clustering Algorithm

In order to develop more accurate teaching plans and methods based on students’ characteristics and needs, and improve the effectiveness of labor education, a two-stage clustering algorithm based method for predicting the effectiveness of labor education in universities is studied. A two-stage clustering algorithm-based method for predicting the effectiveness of labor education in universities has been proposed. Through in-depth analysis of labor education data in universities, principal component analysis (PCA) was used to select 16 most representative variables from the original features. Subsequently, self-organizing map (SOM) neural network was used for preliminary clustering, and the optimal number of clusters was determined to be 8. Furthermore, the K-means clustering algorithm performs fine clustering based on the initial centers provided by SOM, significantly improving the stability and accuracy of clustering. At the same time, processing the outliers in the clustering results has reduced their impact on the clustering effect. Finally, the processed data was input into an improved XGBoost prediction model, which achieved a true positive rate (TPR) of over 85% on the ROC curve when predicting the effectiveness of labor education in universities. While maintaining a low false positive rate (FPR), the model outperformed other comparison methods significantly. This achievement not only validates the effectiveness of the proposed method in feature selection, clustering optimization, and prediction model construction, but also confirms its ability to accurately predict the labor education effectiveness of students in different grades, providing strong support for higher education institutions to optimize teaching plans and improve educational outcomes.

A Method for Evaluating Demand Response Potential of Industrial Loads Based on Fuzzy Control

Demand response (DR) can ensure electricity supply security by shifting or shedding loads, which plays an important role in a power system with a high proportion of renewable energy sources. Industrial loads are vital participants in DR, but it is difficult to assess DR potential because of many complex factors. In this paper, a new method based on fuzzy control is given to assess the DR potential of industrial loads. A complete assessment framework including four steps is presented. Firstly, the industrial load data are preprocessed to mitigate the influence of noisy and transmission losses, and then the K-means algorithm considering the optimal cluster number is used to calculate baseline load of industrial load. Subsequently, an open-loop fuzzy controller is designed to predict the response factor of different industrial loads. Three strongly correlated indicators, namely peak load rate, electricity intensity, and load flexibility, are selected as the input of fuzzy control, which represents response willingness. Finally, the baseline load of diverse clustering scenarios and the response factor are used to calculate the DR potential of different industrial loads. The proposed method takes into account both economic and technical factors comprehensively, and thus, the results better represent the available DR potential in real-world situations. To demonstrate the effectiveness of the proposed method, the case of a medium-sized city in China is studied. The simulation focuses on the top eight industrial types, and the results show they can contribute about 189 MW available DR potential.

Analysis and validation of clinical subgroups of Kawasaki disease in children in China: a retrospective study

ObjectiveAlthough Kawasaki disease (KD) is commonly regarded as a single disease entity, clinical subgroups have recently been described. We aimed to validate previous research on clinical subgroups and establish a...

Financial Stability and Innovation: The Role of Non-Performing Loans

This study analyses the relationship between non-performing loans (NPLs) and innovation systems at a global level. The data were obtained from the World Bank and the Global Innovation Index over the period 2013–2022 for 149 countries. The k-means algorithm was used to verify the presence of clusters in the data. Since k-means is an unsupervised machine-learning algorithm, we compared the Silhouette coefficient with the Elbow method to find an optimization. The results show that the optimal number of clusters is three, as suggested using the Elbow Method. Furthermore, a panel data analysis was conducted. Results show that the level of NPLs is positively associated with cultural and creative services exports as a percentage of total trade and innovation input sub-index and negatively associated with the Hirsch Index, ICT services exports as a percentage of total trade, ICT services imports as a percentage of total trade, and information and communication technologies.

Pengelompokan Provinsi di Indonesia Menggunakan Gaussian Mixture Model Berdasarkan Indikator Kemiskinan

Poverty is a multidimensional concept that can not only be seen from an economic perspective, but can also be seen from a social, cultural and political perspective. Poverty is generally expressed as a static concept with the amount of poverty as a description of welfare conditions at a certain time. In reality, poverty has a time period that is never cut off and has a continuous pattern over time. The Indonesian government is committed to efforts to eradicate poverty in order to reduce Indonesia's poverty rate. One way that can be done is by allocating aid to the community. Grouping is carried out by cluster analysis using the Gaussian Mixture Model method with the Expectation-Maximization (EM) algorithm. The optimal number of clusters in the Gaussian Mixture Model uses the Bayesian Information Criterion (BIC) method. Based on the research carried out, the results of the analysis of cluster grouping using the Gaussian Mixture Model method were obtained with the smallest BIC value, namely -397.6876 with a total of 2 clusters.

IMPLEMENTATION OF K-MEANS AND FUZZY C-MEANS CLUSTERING FOR MAPPING TODDLER STUNTING CASES IN GUNUNGKIDUL DISTRICT

Gunungkidul Regency has the highest prevalence of stunted toddlers in the Special Region of Yogyakarta. This study aims to describe the optimal clustering results of toddler stunting cases using the k-means and fuzzy c-means methods and to describe the characteristic of the mapping results of stunting-prone areas for toddlers in Gunungkidul Regency for the years 2020 – 2022. This study maps stunting-prone areas for toddlers across 30 community health centers in Gunungkidul Regency from 2020 to 2022, with variables including the percentage of babies with low birth weight, babies born stunted, babies receiving health services, stunted toddlers, toddlers receiving health services, babies given exclusive breastfeeding, poor couples of reproductive ages, and families with adequate drinking water. The k-means clustering method determines cluster membership using the distance between objects and centroids, while the fuzzy c-means method uses the degree of membership. Cluster evaluation uses the silhouette coefficient, Calinski-Harabasz index, Davies-Bouldin index, and Dunn index to obtain optimal clustering results. The mapping results are presented as a stunting vulnerability map. The findings indicate that the optimal number of clusters is two, with the fuzzy c-means method proving more optimal than the k-means method based on evaluation scores. In 2020, there were 23 community health centers in cluster 0 and 7 in cluster 1. In 2021, there were 21 community health centers in cluster 0 and 9 in cluster 1. In 2022, there were 18 community health centers in cluster 0 and 12 in cluster 1. Generally, community health centers in cluster 0 are less optimal in specific nutrition interventions, such as for infants and toddlers. In contrast, those in cluster 1 are less optimal in sensitive nutrition interventions, such as poverty and water adequacy.

APPLICATION OF K-MEANS++ WITH DUNN INDEX VALIDATION OF GROUPING WEST KALIMANTAN REGION BASED ON CRIME VULNERABILITY

Crime is an unlawful behavior that will be given a punishment or sanctions based on Kitab Undang-Undang Hukum Pidana (KUHP) or other regulations in Indonesia. One of the provinces in Indonesia, namely West Kalimantan reported that criminal cases are increasing in 2021 and 2022. One of the solutions to minimize that case is grouping the district and city in West Kalimantan based on the level of vulnerability so the authority can be more responsive in solving these problems. The grouping can be done by cluster analysis. This analysis aims to group some objects based on the similarity of characteristics. K-Means++ is one of the methods of cluster analysis. K-Means++ is the development of K-Means, in which K-Means++ is smarter than K-Means in selecting the initial centroid because only one initial centroid is chosen randomly, and the initial centroids of the other clusters are done through calculations. This research uses secondary data from BPS of West Kalimantan, consisting of 10 variables. This research aims to form clusters to determine the level of vulnerability of each district and city in West Kalimantan. The selection of the optimal cluster is done by evaluating the cluster. One of these evaluations is the Dunn Index. Based on the analysis results, the optimum number of clusters is with a Dunn Index value of 0.55. The first cluster is categorized as non-vulnerable with ten members, the second cluster as vulnerable with three members, and the third cluster as very vulnerable with one member.

Identifying Comorbidity Patterns in People with and without Alzheimer's Disease Using Latent Dirichlet Allocation.

Multimorbidity is common in older adults and complicates diagnosing and care for this population. We investigated co-occurrence patterns (clustering) of medical conditions in persons with Alzheimer's disease (AD) and their matched controls. The register-based Medication use and Alzheimer's disease study (MEDALZ) includes 70,718 community-dwelling persons with incident AD diagnosed during 2005-2011 in Finland and a matched comparison cohort. Latent Dirichlet Allocation was used to cluster the comorbidities (ICD-10 diagnosis codes). Modeling was performed separately for AD and control cohorts. We experimented with different numbers of clusters (also known as topics in the field of Natural Language Processing) ranging from five to 20. In both cohorts, 17 of the 20 most frequent diagnoses were the same. Based on a qualitative assessment by medical experts, the cluster patterns were not affected by the number of clusters, but the best interpretability was observed in the 10-cluster model. Quantitative assessment of the optimal number of clusters by log-likelihood estimate did not imply a specific optimal number of clusters. Multidimensional scaling visualized the variability in cluster size and (dis)similarity between the clusters with more overlapping of clusters and variation in group size seen in the AD cohort. Early signs and symptoms of AD were more commonly clustered together in the AD cohort than in the comparison cohort. This study experimented with using natural language processing techniques for clustering patterns from an epidemiological study. From the computed clusters, it was possible to qualitatively identify multimorbidity that differentiates AD cases and controls.