Hierarchical Clustering Method Research Articles

Assessment of PM2.5 Patterns in Malaysia Using the Clustering Method

Particulate matter is the parameter of most concern in air quality monitoring in Malaysia. This study discusses the variations and clustering of PM2.5 recorded from 2018 to 2019 at 65 stations of the Continuous Air Quality Monitoring Network of the Malaysian Department of Environment. PM2.5 concentrations were recorded continuously using a tapered element oscillating microbalance. The cluster analysis was conducted using the Agglomerative Hierarchical Cluster (AHC) method. The results show that the daily average of PM2.5 concentrations ranged between 8 and 31 µg m–3. The cluster regions were classified into High Pollution Regions (HPR), Medium Pollution Regions (MPR) and Low Pollution Regions (LPR) based on the AHC analysis. The mean concentration of PM2.5 recorded in HPR was significantly higher with 23.04 µg m–3 followed by MPR and LPR. The results also showed that the highest concentration of PM2.5 was recorded during the 2019 haze episode for all three regions, with the air pollutant index indicating very unhealthy and dangerous levels.

Clustering Gene Expression Data Through Modified Agglomerative M-CURE Hierarchical Algorithm

Gene expression refers to the process in which the gene information is used in the functional gene product synthesis. They basically encode the proteins which in turn dictate the functionality of the cell. The first step in gene expression study involves the clustering usage. This is due to the reason that biological networks are very complex and the genes volume increases the comprehending challenges along with the data interpretation which itself inhibit vagueness, noise and imprecision. For a biological system to function, the essential cellular molecules must interact with its surrounding including RNA, DNA, metabolites and proteins. Clustering methods will help to expose the structures and the patterns in the original data for taking further decisions. The traditional clustering techniques involve hierarchical, model based, partitioning, density based, grid based and soft clustering methods. Though many of these methods provide a reliable output in clustering, they fail to incorporate huge data of gene expressions. Also, there are statistical issues along with choosing the right method and the choice of dissimilarity matrix when dealing with gene expression data. We propose to use a modified clustering algorithm using representatives (M-CURE) in this work which is more robust to outliers as compared to K-means clustering and also able to find clusters with size variances.

Методика визначення та розташування оптимальної кількості зарядних станцій для електротранспорту в населеному пункті

The problem of determining the optimal number and location of charging stations for electric transport within the settlement is being studied. To solve this problem, the criteria, factors and limitations of the optimal placement of charging stations were defined. Since the task included ensuring the comfortable operation of electric cars by citizens until 2025, it was necessary to forecast the number of electric cars that will be registered in the town of Cherkasy. For this, a portrait of the potential consumer was drawn up, and the potential capacity of the transport market was determined. Also, the key points of concentration of potential customers were determined, the level of charging stations for electric vehicles that will ensure their operation in the town of Cherkasy was selected, and the calculation of the optimal number of charging stations for the South-Western district of Cherkasy was performed. Since there are no regulations regarding the required number of charging stations for a certain fleet, it was proposed to adapt the regulations for gas stations by making corrections regarding the features of the technical charging process and the number of electric vehicle charges per unit of time. With the help of the method of hierarchical clustering, demand points in charging stations for electric cars were determined and a map of their location was developed for the residents of the South-Western district of Cherkasy. The cost of this project was also calculated. It can be concluded that the implementation of the project to provide charging stations for electric cars is profitable both from the point of view of material investments and from the point of view of the occupied useful area of town parking lots. The administration of Cherkasy may be interested in the implementation of this project under the terms of the city development program. This is due to the fact that the implementation of this project is expected to improve the town's environmental situation, generate profit from charging electric cars and sell new electric cars at car dealerships.

Dissolution of Mn-bearing dolomite drives elevated Cr(VI) occurrence in a Permian redbed aquifer.

Municipalities in central Oklahoma, U.S.A. increasingly rely on water drawn from the Central Oklahoma Aquifer (COA) as surface water resources have not grown in proportion to population and current water demands. However, water drawn from certain regions of the COA frequently contains elevated levels of naturally occurring hexavalent chromium. Rock samples from the Norman Arsenic Test Hole Core (NATHC) were investigated to identify the mineralogic host(s) of Cr and mechanisms of Cr(VI) release via bulk mineralogy and chemistry measurements, selective chemical extractions, and microscale elemental analyses. Results demonstrate most COA Cr is contained in Fe oxides and clays as isomorphic substitutions for Fe(III). Analyses of regional groundwater data, including hierarchical clustering methods and GIS, demonstrate the most intense Cr(VI) occurrence is linked to cation exchange with Na-clays at depth. Cation exchange allows dissolution of Mn-bearing dolomite, which in turn produces Mn oxides in otherwise dolomite-saturated groundwaters. Mn oxides in turn are known to oxidize Cr(III) to Cr(VI). In general, co-occurrence of Mn-bearing carbonates and exchangeable clays in any aquifer, particularly those with Cr(III) present in iron oxide cements, serve as ingredients for groundwater occurrences of oxidizable trace metals.

Grasshopper KUWAHARA and Gradient Boosting Tree for Optimal Features Classifications

This paper aims to design an optimizer followed by a Kawahara filter for optimal classification and prediction of employees’ performance. The algorithm starts by processing data by a modified K-means technique as a hierarchical clustering method to quickly obtain the best features of employees to reach their best performance. The work of this paper consists of two parts. The first part is based on collecting data of employees to calculate and illustrate the performance of each employee. The second part is based on the classification and prediction techniques of the employee performance. This model is designed to help companies in their decisions about the employees’ performance. The classification and prediction algorithms use the Gradient Boosting Tree classifier to classify and predict the features. Results of the paper give the percentage of employees which are expected to leave the company after predicting their performance for the coming years. Results also show that the Grasshopper Optimization, followed by “KF” with the Gradient Boosting Tree as classifier and predictor, is characterized by a high accuracy. The proposed algorithm is compared with other known techniques where our results are fund to be superior.

Features of the Hospitalized Incidence and Prevalence of Pediatric and Adolescent Endocrine Disorders in the Orenburg Region

Background: The incidence of endocrine diseases, eating and metabolic disorders has been increasing recently. The prevalence and structure of pediatric and adolescent endocrine disorders are significantly different from those in adults and have specific regional features. Objective: To study the frequency, structure, and long-term changes in the rates of hospitalization of the child population for diseases of the endocrine system in the Orenburg Region and to assess local features of disease incidence and prevalence. Materials and methods: We analyzed data on all pediatric cases (aged 0–18) of hospital care for endocrine disorders provided by the regional office of the Federal Compulsory Medical Insurance Fund for 2016–2020. We also applied the Ward’s hierarchical clustering method to establish local peculiarities and to group municipalities by the rates and dynamics of endocrine diseases in children. Results: The average rate of hospitalization for endocrine disorders was 3.8 cases per 1,000 children and accounted for almost 3.9 % of all types of hospital care. The proportion of inpatients among children aged 0 to 14 years was higher than that among adolescents. Metabolic disorders (34.0 %), insulin-dependent diabetes mellitus (22.3 %), hypofunction and other pituitary disorders (12.6 %), obesity and other types of overnutrition (7.1 %), and undernutrition (5.7 %) prevailed in all hospitalized children. The patterns of pediatric and adolescent inpatients differed significantly. The analysis of the prevalence of endocrine disorders in children and its long-term changes enabled clustering of the territories of the Orenburg Region.Conclusions: Three distinct clusters of territories were established in the analysis of local rates of endocrine diseases in children. The first cluster was characterized by moderate disease rates, while the second and the third clusters demonstrated high and low incidence and prevalence rates of pediatric and adolescent endocrine disorders.

Performance Improvement of Very Short-term Prediction Intervals for Regional Wind Power Based on Composite Conditional Nonlinear Quantile Regression

Accurate regional wind power prediction plays an important role in the security and reliability of power systems. For the performance improvement of very short-term prediction intervals (PIs), a novel probabilistic prediction method based on composite conditional nonlinear quantile regression (CCNQR) is proposed. First, the hierarchical clustering method based on weighted multivariate time series motifs (WMTSM) is studied to consider the static difference, dynamic difference, and meteorological difference of wind power time series. Then, the correlations are used as sample weights for the conditional linear programming (CLP) of CCNQR. To optimize the performance of PIs, a composite evaluation including the accuracy of PI coverage probability (PICP), the average width (AW), and the offsets of points outside PIs (OPOPI) is used to quantify the appropriate upper and lower bounds. Moreover, the adaptive boundary quantiles (ABQs) are quantified for the optimal performance of PIs. Finally, based on the real wind farm data, the superiority of the proposed method is verified by adequate comparisons with the conventional methods.

Using the Agglomerative Hierarchical Clustering Method to Examine Human Factors in Indonesian Aviation Accidents

This study aims to provide a comprehensive source of knowledge regarding aviation accidents in Indonesia caused by human factors, which are the most significant among other causative elements, requiring a detailed assessment of the accident as a result of pilot and co-pilot faults while operating the aircraft. The KNKT website database is still in the form of accident reports. To this end, the retrieved information based on historical data for 23 years of accidents caused by humans by analyzing the data using the clustering approach to gain data insight in the relationship between total flying hours and pilot licenses. The data analysis revealed that, in general, the aircraft operator complied with the CASR standards.

Extracting hierarchical features of cultural variation using network-based clustering.

High-dimensional datasets on cultural characters contribute to uncovering insights about factors that influence cultural evolution. Because cultural variation in part reflects descent processes with a hierarchical structure – including the descent of populations and vertical transmission of cultural traits – methods designed for hierarchically structured data have potential to find applications in the analysis of cultural variation. We adapt a network-based hierarchical clustering method for use in analysing cultural variation. Given a set of entities, the method constructs a similarity network, hierarchically depicting community structure among them. We illustrate the approach using four datasets: pronunciation variation in the US mid-Atlantic region, folklore variation in worldwide cultures, phonemic variation across worldwide languages and temporal variation in first names in the US. In these examples, the method provides insights into processes that affect cultural variation, uncovering geographic and other influences on observed patterns and cultural characters that make important contributions to them.

Enhancement of Design Level Class Decomposition using Evaluation Process

Refactoring on the design level artifact such as the class diagram was already done using the threshold-based agglomerative hierarchical clustering method, specifically class decomposition. The approach produced a better cluster based on the label name similarity of attribute and method. But, some problems emerge from the experiment result. The negative Silhouettes element still exist in the cluster. And, there is an unusable cluster that only consists of one attribute element. This paper has proposed the evaluation process to optimize the result of clustering. This evaluation process is an additional process that aims to move the negative Silhouettes element to the other cluster. The movement is also to get the better value of element Silhouettes value. The evaluation process can produce a better result for clusters. The clusters produced from the evaluation process have higher Silhouettes values. The average Silhouettes value is increased by about 40%. Ultimately, the result shows no unusable cluster as mentioned in the previous research.

Does non-linear factorization of financial returns help build better and stabler portfolios?

A portfolio allocation method based on linear and non-linear latent constrained conditional factors is presented. The factor loadings are constrained to always be positive in order to obtain long-only portfolios, which is not guaranteed by classical factor analysis or PCA. In addition, the factors are to be uncorrelated among clusters in order to build long-only portfolios. Our approach is based on modern machine learning tools: convex Non-negative Matrix Factorization (NMF) and autoencoder neural networks, designed in a specific manner to enforce the learning of useful hidden data structure such as correlation between the assets' returns. Our technique finds lowly correlated linear and non-linear conditional latent factors which are used to build outperforming global portfolios consisting of cryptocurrencies and traditional assets, similar to hierarchical clustering method. We study the dynamics of the derived non-linear factors in order to forecast tail losses of the portfolios and thus build more stable ones.

Hierarchical clustering analysis of hydrogen bond networks in aqueous solutions.

To understand the relation between the macroscopic properties and microscopic structure of hydrogen bond networks in solutions, we introduced a hierarchical clustering method to analyze the typical configurations of water clusters in this type of network. Regarding hydrogen bonds as frames, the rings, fragments and clusters are defined and analyzed to provide a comprehensive perspective for the distributional and dynamic characteristics of the hydrogen-bonding network in NaCl solution at different concentrations. The properties of the radial distribution function and hydrogen bonds are first analyzed. Destruction and shorter lifetimes of hydrogen bonds are observed in solutions. In further analysis of the two-dimensional configuration, i.e., ring, and three-dimensional configuration, i.e., fragment, the average number, size and lifetime of these structures consistently decrease as the concentration increases. Ionic effects on disrupting rings and fragments are significant in the first hydration shell, especially with sodium cations, and these effects weaken beyond the first hydration shell. Regarding the clusters obtained using the Louvain algorithm, our results indicate that clusters break and become smaller as the NaCl concentration increases. The presence of ions also leads to the isolation of clusters and therefore the inhibition of changes. The lifetime of clusters increases with NaCl concentration, indicating the slowed breakage and reformation of clusters in NaCl solutions. This method can be further applied to quantitatively characterize hydrogen bond networks to elucidate more properties of aqueous solutions.

Clustering Approach to Generate Pedestrian Traffic Pattern Groups


 
 
 
 This study shows the development of patterns of temporal hourly volume distributions in an urban area in Costa Rica, based on a cluster analysis of pedestrian data. This study aims to establish specific pattern groups for the temporal variation of weekday pedestrian volumes applying cluster analysis in the central business district of Guadalupe in San José. For 46 counting sites, vectors with the weekday hourly factors, the proportion of the daily pedestrian traffic, were estimated. A hierarchical cluster method was implemented to group the vectors of hourly factors from the different counting sites. This method groups elements by minimizing the Euclidean distance between elements of the same group and, at the same time, maximizing the distances from elements of other groups. In addition, the groups found through this analysis are related to land use through buffers of different radios. Eight temporal pattern groups were obtained through cluster analysis. Two pattern groups account for more than two-thirds of the sites included in the study. Fisher’s exact independence test shows that banks and public services could explain some of the patterns observed. The classification of 46 counting sites based on temporal distribution patterns, and the relation with the establishments in the area, allows a simplification of the information and facilitates an understanding of the pedestrian mobility in the area. Further research is required that leads towards geographical elements that could explain the differences in temporal and mobility patterns.
 
 
 

A cluster analysis of basketball players for each of the five traditionally defined positions

Determining the players’ playing styles and bringing the right players together are very important for winning in basketball. This study aimed to group basketball players into similar clusters according to their playing styles for each of the traditionally defined five positions (point guard (PG), shooting guard (SG), small forward (SF), power forward (PF), and center (C)). This way, teams would be able to identify their type of players to help them determine what type of players they should recruit to build a better team. The 17 game-related statistics from 15 seasons of the National Basketball Association (NBA) were analyzed using a hierarchical clustering method. The cluster validity indices (CVIs) were used to determine the optimum number of groups. Based on this analysis, four clusters were identified for PG, SG, and SF positions, while five clusters for PF position and six clusters for C position were established. In addition to the definition of the created clusters, their individual achievements were examined based on three performance indicators: adjusted plus-minus (APM), average points differential, and the percentage of clusters on winning teams. This study contributes to the evaluation of team compatibility, which is a significant part of winning, as it allows one to determine the playing styles for each position, while examining the success of position pair combinations.

Study on the Influence of Physicochemical Indexes on Wine Quality

The quality of a wine is generally determined by hiring a group of qualified wine judges to taste the wine. The quality of wine grapes is directly related to the quality of wine. To explore how the physical and chemical indicators of wine and wine grapes can reflect the quality of wine and grape to a certain extent. This article is based on 2012 Chinese college students' mathematical contest in modeling A problem of data, to score as the dependent variable, wine physical and chemical indicators as independent variables, the method of using principal component analysis of physical and chemical indexes of wine grape dimension reduction, physical and chemical indexes selection of grapes, and use the decision tree regression method to establish score and wine function relation between the physical and chemical indicators. Then, after the comparison of the results of the four clustering algorithms, the hierarchical clustering method based on Agnes algorithm was used to conduct cluster analysis on the wines. Considering the rationality of classification, red and white wines were divided into four grades, and the average value of the estimated scores of each grade was used as the benchmark score for the wines of this grade.

A Cluster Analysis Concerning the Behavior of Enterprises with E-Commerce Activity in the Context of the COVID-19 Pandemic

The planning of activities of e-commerce enterprises and their behavior has been influenced by the emergence of the COVID-19 pandemic. The behavior of e-commerce enterprises has been highlighted at the level of EU countries through an analysis elaborated on four variables: the value of e-commerce sales, cloud computing services, enterprises that have provided training to develop/upgrade the ICT skills of their personnel, e-commerce, customer relationship management (CRM) and secure transactions. Using the hierarchical clustering method, analysis was carried out on these variables to identify certain economic and behavioral patterns of e-commerce activity from 2018 and 2020. The study of the relationships involved in the e-commerce activity of these enterprises is reflected in models of the economic behavior of 31 European states in relation to the targeted variables. The results show that the impacts of the COVID-19 pandemic are strongly manifested in the direction of the evolution of each indicator but differ from one country to another. The trends depend on the level of development and the particularities of each country’s economy in adapting to the repercussions reported in relation to the level of impact of the COVID-19 pandemic. This is highlighted by the significant regrouping of countries in 2020 compared with 2018 in relation to the average values of the indicators. The results show that, in 2020, the most significant percentages of the value of e-commerce sales were recorded in Belgium, Ireland and Czechia, as in 2018. In e-commerce, customer relationship management and secure transactions, Denmark and Sweden were superior in 2020 to the countries mentioned above, which were dominant in 2018. For the other two indicators, Finland and Norway were the top countries included in the analysis in both years. The conclusion supports the continuous model of e-commerce enterprise behavior in order to meet the requirements of online customers.

Performance Comparison with Hierarchical and Partitional Clustering Methods

In data mining, one of the commonly-used techniques is the clustering. Clustering can be done by the different algorithms such as hierarchical, partitioning, grid, density and graph based algorithms. In this study first of all the concept of data mining explained, then giving information the aims of using data mining and the areas of using and then clustering and clustering algorithms that used in data mining are explained theoretically. Ultimately within the scope of this study, "Mall Customers" data set that taken from Kaggle database, based partitioned clustering and hierarchical clustering algorithms aimed at the separation of clusters according to their costumers features. In the clusters obtained by the partitional clustering algorithms, the similarity within the cluster is maximum and the similarity between the clusters is minimum. The hierarchical clustering algorithms is based on the gathering of similar features or vice versa. The partitional clustering algorithms used; k-means and PAM, hierarchical clustering algorithms used; AGNES and DIANA are algorithms. In this study, R statistical programming language was used in the application of algorithms. At the end of the study, the data set was run with clustering algorithms and the obtained analysis results were interpreted.

Symplectic Ramanujan Mode Decomposition and its application to compound fault diagnosis of bearings

The existing compound-fault diagnosis methods of rolling bearings have their own defects, which makes their accuracy of fault diagnosis impossible to be guaranteed. Therefore, this paper attempts to combine symplectic similarity transformation with Ramanujan subspace theory, and then a periodic impulse extraction method called symplectic Ramanujan mode decomposition (SRMD) method is proposed. SRMD separates the components with different fault features through symplectic similarity transformation and hierarchical clustering method to obtain symplectic clustering components (SCCs). At the same time, SRMD uses the Ramanujan subspace theory to extract the major periodic impulse components of each component to be extracted, and then obtains symplectic Ramanujan components (SRCs). The results show that SRMD is a resultful method in compound-fault diagnosis of bearings with excellent periodic impulse extraction ability.

Availability and Suitability of Digital Health Tools in Africa for Pandemic Control: Scoping Review and Cluster Analysis.

BackgroundGaining oversight into the rapidly growing number of mobile health tools for surveillance or outbreak management in Africa has become a challenge.ObjectiveThe aim of this study is to map the functional portfolio of mobile health tools used for surveillance or outbreak management of communicable diseases in Africa.MethodsWe conducted a scoping review by combining data from a systematic review of the literature and a telephone survey of experts. We applied the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines by searching for articles published between January 2010 and December 2020. In addition, we used the respondent-driven sampling method and conducted a telephone survey from October 2019 to February 2020 among representatives from national public health institutes from all African countries. We combined the findings and used a hierarchical clustering method to group the tools based on their functionalities (attributes).ResultsWe identified 30 tools from 1914 publications and 45 responses from 52% (28/54) of African countries. Approximately 13% of the tools (4/30; Surveillance Outbreak Response Management and Analysis System, Go.Data, CommCare, and District Health Information Software 2) covered 93% (14/15) of the identified attributes. Of the 30 tools, 17 (59%) tools managed health event data, 20 (67%) managed case-based data, and 28 (97%) offered a dashboard. Clustering identified 2 exceptional attributes for outbreak management, namely contact follow-up (offered by 8/30, 27%, of the tools) and transmission network visualization (offered by Surveillance Outbreak Response Management and Analysis System and Go.Data).ConclusionsThere is a large range of tools in use; however, most of them do not offer a comprehensive set of attributes, resulting in the need for public health workers having to use multiple tools in parallel. Only 13% (4/30) of the tools cover most of the attributes, including those most relevant for response to the COVID-19 pandemic, such as laboratory interface, contact follow-up, and transmission network visualization.

Diversity of the Seed Material of Selected Plant Species of Naturally Valuable Grassland Habitats in Terms of the Prognosis of Introduction Success

The current conservation status of semi-natural grassland habitats in Poland indicates that actions for their restoration are necessary. Many of the degraded sites require the introduction of diaspores of representative species because other methods of improving their condition are not sufficiently effective. Meanwhile, little is known about the diaspores of native wild-flower species and the biology of their seeds. The aim of the present study was to find an answer to the question of which features of the seed material can guarantee the success of the introduction. The study covered 28 plant species of 4 non-forest natural habitats (codes: 6440, 6410, 6510, 6210) occurring in river valleys. Diaspores were collected in 2015–2017. Morphometric measurements of diaspores were performed, the weight of 1000 diaspores was determined and the germination capacity analysis was carried out in accordance with ISTA Rules. The analysis was made with the division into normal seedlings, abnormal seedlings, dead seeds and fresh ungerminated seeds. Species with similar parameters of seed material were selected using the method of hierarchical clustering and PCA analysis. Three groups of species were distinguished: (1) with good seed germination capacity (above 65%), which, regardless of weather conditions during the generative development, and despite the small size of diaspores, can guarantee successful introductions (Verbascum thapsus, Veronica longifolia, Daucus carota, Plantago lanceolata); (2) species of little suitability for introduction, due to the large proportion (over 50%) of dead seeds (Armeria maritima, Linaria vulgaris, Potentilla erecta, Centaurea stoebe, Sanguisorba officinalis, Cnidium dubium); (3) species with relatively large size of diaspores and low seed germination capacity, due to the high proportion of fresh ungerminated seeds which means dormant seeds (Lathyrus pratensis, Geranium pratense).