Density-based Clustering Algorithm Research Articles

Klasifikasi Penggunaan Teknologi Pada Petani Milenial di Sulawesi Selatan Menggunakan Density Based Spatial Clustering Algorithm With Noise

This research examines the use of agricultural technology by millennial farmers in South Sulawesi Province by applying the Density-Based Spatial Clustering Algorithm with Noise (DBSCAN) to the agricultural census data of South Sulawesi Province Phase 1 in 2023. The secondary data used includes the number of millennial farmers aged 19-39 who use or do not use digital technology, divided by district/city and gender. The analysis process begins with preprocessing to prepare the data, followed by clustering using the DBSCAN algorithm, determining the optimal values for the Eps and minPts parameters, and evaluating the quality of the formed clusters using the silhouette coefficient and elbow method. The results of the study indicate that the combination of Eps value of 1.000 and minPts value of 7 produces optimal clustering with 2 clusters formed and 92 data points clustered, while 4 other data points are considered as noise. Evaluation using the silhouette coefficient and elbow method also indicates that the optimal data grouping is k=2.

A scalable multi-density clustering approach to detect city hotspots in a smart city

In the field of Smart City applications, the analysis of urban data to detect city hotspots, i.e., regions where urban events (such as pollution peaks, virus infections, traffic spikes, and crimes) occur at a higher density than in the rest of the dataset, is becoming a common task. The detection of such hotspots can serve as a valuable organizational technique for framing detailed information about a metropolitan area, providing high-level spatial knowledge for planners, scientists, and policymakers. From the algorithmic viewpoint, classic density-based clustering algorithms are very effective in discovering hotspots characterized by homogeneous density; however, their application on multi-density data can produce inaccurate results. For such a reason, since metropolitan cities are characterized by areas with significantly variable densities, multi-density clustering approaches are more effective in discovering city hotspots. Moreover, the growing volumes of data collected in urban environments require the development of parallel approaches, in order to take advantage of scalable executions offered by Edge and Cloud environments. This paper describes the design and implementation of a parallel multi-density clustering algorithm aimed at analyzing high volumes of urban data in an efficient way. The experimental evaluation shows that the proposed parallel clustering approach takes out encouraging advantages in terms of execution time, speedup, and efficiency.

Performance Analysis and Evaluation of Clustering Algorithms using WEKA

Clustering, an unsupervised learning technique, to find inherent groupings in un-labelled data. It seems to be referring to a study or research paper that examines and uses a number of clustering algorithms, including the canopy method, k-Means clustering, hierarchical clustering, density-based clustering, and EM algorithm. WEKA, a clustering program, is used for the examination of these techniques. and the effectiveness of these algorithms is evaluated through experiments using social network Ads datasets. The goal of this research paper or study seems to be to assess how well these clustering algorithms perform in grouping data within social network Ads datasets. Such analyses can help identify the most suitable algorithm for a specific type of data or problem domain and may lead to insights into the underlying structure of the data.

Automatic Group Decision-Making for Algal Bloom Management Based on Information Self-Learning

The phenomenon of algal bloom seriously affects the function of the aquatic ecosystems, damages the landscape of urban river and lakes, and threatens the safety of water use. The introduction of a multi-attribute decision-making method avoids the shortcomings of traditional algal bloom management that relies on manual experience. However, the weight-calculation part of this method still receives the artificial influence of human factors, which reduces the accuracy and scientific rigor of the decision. This paper presents a group decision-making method based on information self-learning which makes decision weights automatically clustered and assigned. A general framework of decision-making management is constructed for the algal bloom management process. In the decision-making process, an improved density-based clustering algorithm is used to automatically cluster and rank the decision data in the form of the three-parameter interval number, and ultimately obtain the optimal management method that meets the management objectives. Finally, the method was applied at the monitoring station of Sanjiadian Reservoir in Beijing, China. Based on the treatment objectives and water quality monitoring data of the station, relevant experts were invited to evaluate the management solutions, and the information self-organizing algorithm of this paper was used to automatically rank the decision-making methods, and finally obtain the most suitable management method for the station. Comparison with the water quality data and treatment inputs after the previous man-made selection of treatment options, and discussion among experts, show that the decision-making method is feasible and effective, and contributes to the sustainable treatment of algal blooms.

A framework for ship abnormal behaviour detection and classification using AIS data

This paper proposes a method for detecting and classifying ship abnormal behaviour in ship trajectories. The method involves generating parameter profiles for the ship's trajectory and applying a Sliding Window algorithm to detect the ship's abnormal behaviour. Then, several features are adopted to effectively describe the characteristics of each ship's abnormal behaviour, such as the standard deviation of speed, detour factor, maximum drift angle, accumulative change of Course Over Ground and maximum lateral distance to the ship route. A density-based clustering algorithm is applied to group similar abnormal behaviour patterns according to the feature similarity, and the Random Forest Classification method is used to train a classification model based on the features extracted from the clusters. The proposed method is then tested on historical ship trajectory data provided by the Automatic Identification System. The results suggest that the method effectively identifies and classifies different abnormal behaviours in the ship trajectories.

Observed links between heatwaves and wildfires across Northern high latitudes

Data on Arctic and Sub-Arctic summer heat events are limited due to the sparse network of surface observation stations. Here, we analyze large heat events within 60°–80°N using land surface temperature (LST) observations from the moderate resolution imaging spectroradiometer (MODIS) sensor aboard the Terra satellite. Our heatwave (HW) detection method uses exceedances of the climatological 90th percentile of LST across summer months, and a spatio-temporal density-based clustering algorithm to distinguish space-time coherent events across Northern Hemispheric high latitudes for the summers of 2000–2022. We find a close link between HW duration, spatial extent and amplitude across the study region (correlations ranging from 0.63 to 0.73). MODIS-derived burned area data show that wildfire seasons are significantly correlated to summer HW activity, particularly in Siberia (r = 0.87 at p < 0.05) and Alaska and NW Canada (r = 0.45 at p < 0.05), and are also spatially co-located. MODIS active fire data also show substantial increase during larger heat events. For the strongest HWs in Siberia, the peak in daily fire count (from the MODIS active fire archive) lags behind peak HW activity. We conclude that there is a close link between intense fire summers and extensive HWs over boreal and shrub tundra regions (Interior Alaska, the Canadian Prairies and Eastern Siberia).

Subgroups of High-cost Patients and Their Preventable Inpatient Cost in Rural China.

High-cost patients account for most healthcare costs and are highly heterogeneous. This study aims to classify high-cost patients into clinically homogeneous subgroups, describe healthcare utilization patterns of subgroups, and identify subgroups with relatively high preventable inpatient cost (PIC) in rural China. A population-based retrospective study was performed using claims data in Xi County, Henan Province. 32,108 high-cost patients, representing the top 10% of individuals with the highest total spending, were identified. A density-based clustering algorithm combined with expert opinions were used to group high-cost patients. Healthcare utilization (including admissions, length of stay and outpatient visits) and spending characteristics (including total spending, and the proportion of PIC, inpatient and out-of-pocket spending on total spending) were described among subgroups. PIC was calculated based on potentially preventable hospitalizations which were identified according to the Agency for Healthcare Research and Quality Prevention Quality Indicators algorithm. High-cost patients were more likely to be older (M=51.87, SD=22.28), male (49.03%) and from poverty-stricken families (37.67%) than non-high-cost patients, with 2.49 (SD=2.47) admissions and 3.25 (SD=4.52) outpatient visits annually. Fourteen subgroups of high-cost patients were identified: chronic disease, non-trauma diseases which need surgery, female disease, cancer, eye disease, respiratory infection/inflammation, skin disease, fracture, liver disease, vertigo syndrome and cerebral infarction, mental disease, arthritis, renal failure, other neurological disorders. The annual admissions ranged from 1.83 (SD=1.23, fracture) to 12.21 (SD=9.26, renal failure), and the average length of stay ranged from 6.61 (SD=10.00, eye disease) to 32.11 (SD=28.78, mental disease) days among subgroups. The chronic disease subgroup showed the largest proportion of PIC on total spending (10.57%). High-cost patients were classified into 14 clinically distinct subgroups which had different healthcare utilization and spending characteristics. Different targeted strategies may be needed for subgroups to reduce preventable hospitalizations. Priority should be given to high-cost patients with chronic diseases.

Exploring Tourist Flow Patterns through Geotagged Social Media Data: A Case Study from Sri Lanka

Tourism studies often rely on conventional methods such as interviews and site-specific surveys to collect data on tourist behaviour, mobility patterns, and preferences. However, these methods can be expensive, time-consuming, and limited in scope. Additionally, accurately tracing tourist travel paths can be challenging due to recall bias. Recently, user generated content on social media platforms have emerged as an alternative data source in tourism studies. In this study, we used geotagged posts on Flickr to understand the dominant paths taken by inbound tourists in Sri Lanka. The methodology consisted of three steps. First, geotagged photographs were collected from the Flickr API spanning a ten-year period, from 2012 to 2022. These photographs included metadata such as user ID, the timestamp of photo capture, and geo coordinates indicating where the photo was taken. Second, a density-based clustering algorithm was utilized to identify tourist hotspots. Finally, the Markov Chain model was employed to calculate transition probabilities among different attractions, revealing dominant travel routes within the country. Study’s findings indicate that Cultural heritage attractions were the most popular, comprising 55% of all attractions identified by the algorithm, and are particularly popular in districts such as Anuradhapura, Polonnaruwa, Matale, Kandy, and Galle. Nature-based attractions, constituting 37% of the total, were mainly located in Nuwara Eliya and Badulla Districts, as well as Yala National Park in Hambantota District. For coastal tourism, Galle and Matara Districts were the top preferences. The analysis confirms dominant travel patterns between North Central, Central, and Southwest coast regions. Notably, Galle, Kandy, Matale, Matara, Nuwara Eliya, and Colombo attract tourists from a more diverse array of regions compared to other districts in the country, showcasing their significance as tourist hubs. The study findings highlight opportunities in Jaffna for cultural-heritage tourism, the Eastern coast for coastal tourism, and the central highlands for tea tourism. It also emphasizes the need to develop the less-visited natural sites to ease pressure on popular National Parks. This research's significance lies in its contribution to informed decision-making and the sustainable management of tourist destinations. By understanding tourist mobility patterns and identifying popular attractions through social media data, policymakers can effectively manipulate visitor flows and mitigate excessive tourist pressure, while preserving the authenticity and allure of these destinations.&#x0D;  &#x0D; Keywords: Sustainable tourism, Geotagged photographs, Social media, Tourist mobility, Flickr

Advanced Seismic Data Analysis: Comparative study of Machine Learning and Deep Learning for Data Prediction and Understanding

This study delves into the application of machine learning (ML) and deep learning (DL) techniques for the analysis of seismic data, aiming to identify and categorize patterns and anomalies within seismic events. Using a robust dataset, we applied three distinct clustering approaches: K-Means, DBSCAN, and an Autoencoder-based method, each offering unique perspectives on the data. K-Means clustering provided a fundamental partitioning of the data into five predefined clusters, facilitating the identification of broad seismic patterns. DBSCAN, a density-based clustering algorithm, offered insights into the spatial distribution and density of seismic events, adeptly pinpointing anomalies and outliers that signify unusual seismic activity. The Autoencoder, leveraging deep learning, excelled in capturing complex and non-linear relationships within the data, revealing subtle patterns not immediately apparent through traditional methods. The effectiveness of these clustering techniques was quantitatively evaluated using the Silhouette Score and the Davies-Bouldin Score, alongside visual assessments through PCA and t-SNE for dimensionality reduction. The results indicated that while K-Means provided clear partitioning, DBSCAN excelled in outlier detection, and the Autoencoder offered a balanced approach with its nuanced analysis capabilities. Our comprehensive analysis underscores the significance of employing a multi-methodological approach in seismic data analysis, as each method contributes uniquely to the understanding of seismic events. The insights gained from this study are valuable for enhancing predictive models and improving disaster risk management strategies in seismology. Future research directions include the integration of additional seismic features, validation against larger datasets, and the development of hybrid models to further refine the predictive accuracy of seismic event analysis.

Individual In-Situ GPS-Derived Acceleration-Speed Profiling: Toward Automatization and Refinement in Male Professional Rugby Union Players

BackgroundRecently a proof-of-concept was proposed to derive the soccer players’ individual in-situ acceleration-speed (AS) profile from global positioning system (GPS) data collected over several sessions and games. The present study aimed to propose an automatized method of individual GPS-derived in-situ AS profiling in a professional rugby union setting.MethodAS profiles of forty-nine male professional rugby union players representing 61.5 million positions, from which acceleration was derived from speed during 51 training sessions and 11 official games, were analyzed. A density-based clustering algorithm was applied to identify outlier points. Multiple AS linear relationships were modeled for each player and session, generating numerous theoretical maximal acceleration (A0), theoretical maximal running speed (S0) and AS slope (ASslope, i.e., overall orientation of the AS profile). Each average provides information on the most relevant value while the standard deviation denotes the method accuracy. In order to assess the reliability of the AS profile within the data collection period, data were compared over two 2-week phases by the inter-class correlation coefficient. A0 and S0 between positions and type of sessions (trainings and games) were compared using ANOVA and post hoc tests when the significant threshold had been reached.ResultsAll AS individual profiles show linear trends with high coefficient of determination (r2 > 0.81). Good reliability (Inter-class Correlation Coefficient ranging from 0.92 to 0.72) was observed between AS profiles, when determined 2 weeks apart for each player. AS profiles depend on players’ positions, types of training and games. Training and games data highlight that highest A0 are obtained during games, while greatest S0 are attained during speed sessions.ConclusionsThis study provides individual in-situ GPS-derived AS profiles with automatization capability. The method calculates an error of measurement for A0 and S0, of paramount importance in order to improve their daily use. The AS profile differences between training, games and playing positions open several perspectives for performance testing, training monitoring, injury prevention and return-to-sport sequences in professional rugby union, with possible transferability to other sprint-based sports.Key PointsAS profiles computed from rugby union GPS data provide positional benchmarks during training and competition.This study provides automatic detection of atypical data and the computation of error measurement of theoretical maximal acceleration and speed components.This refinement constitutes a step forward for a daily use of ecological data by considering data collection and method reliabilities.This easy-to-implement approach may facilitate its use to the performance management process (talent identification, training monitoring and individualization, return-to-sport).

UP-DPC: Ultra-scalable parallel density peak clustering

Density Peak Clustering (DPC) is a highly effective density-based clustering algorithm, but its scalability is limited by the expensive Density Peak Estimation (DPE) step. To address this challenge, we propose UP-DPC: Ultra-Scalable Parallel Density Peak Clustering, a novel framework that employs approximate Density Peak Estimation and performs DPC on LDP-wise graphs. This approach enables UP-DPC to handle datasets of arbitrary scale without relying on spatial indexing for acceleration. Furthermore, we introduce a five-layer computational architecture and leverage parallel computation techniques to further enhance the speed and efficiency of UP-DPC.To evaluate the scalability and effectiveness of UP-DPC, we conduct extensive experiments on 14 datasets, including the large/web-scale datasets, and compare UP-DPC with 21 algorithms. Notably, on the MNIST8M dataset consisting of 8,000k data objects, UP-DPC achieves an NMI (Normalized Mutual Information) value of 0.6464 in just 35.41 seconds, outperforming the state-of-the-art GPU-based method, which only archives an NMI of 0.045 in 56.96 seconds. These results demonstrate the superior scalability and effectiveness of UP-DPC in handling large/web-scale datasets. The proposed framework offers significant improvements over existing methods and shows promise as a solution for density-based clustering tasks.

Online Sparse Representation Clustering for Evolving Data Streams.

Data stream clustering can be performed to discover the patterns underlying continuously arriving sequences of data. A number of data stream clustering algorithms for finding clusters in arbitrary shapes and handling outliers, such as density-based clustering algorithms, have been proposed. However, these algorithms are often limited in their ability to construct and merge microclusters by measuring the Euclidean distances between high-dimensional data objects, e.g., transferring valuable knowledge from historical landmark windows to the current landmark window, and exploiting evolving subspace structures adaptively. We propose an online sparse representation clustering (OSRC) method to learn an affinity matrix for evaluating the relationships among high-dimensional data objects in evolving data streams. We first introduce a low-dimensional projection (LDP) into sparse representation to adaptively reduce the potential negative influence associated with the noise and redundancy contained in high-dimensional data. Then, we take advantage of the l2,1 -norm optimization technique to choose the appropriate number of representative data objects and form a specific dictionary for sparse representation. The specific dictionary is integrated into sparse representation to adaptively exploit the evolving subspace structures of the high-dimensional data objects. Moreover, the data object representatives from the current landmark window can transfer valuable knowledge to the next landmark window. The experimental results based on a synthetic dataset and six benchmark datasets validate the effectiveness of the proposed method compared to that of state-of-the-art methods for data stream clustering.

Vision based road profile estimation for preview-controlled vehicle suspension systems

In this paper, a vision-based road profile estimation method was studied for the control of semi-active and active suspension systems. For the purpose, a monocular camera was used to collect data from the road tests to develop a logic to convert the camera measurements into the road profile data. For the generation of the road profile, alignment of the different sets of camera measurements and their coherence were expressed. Importance of the sensor and process noise removal were shown in recognition of the high frequency content of the road profile, which was a particular interest of the study. Additionally, a density-based clustering algorithm was taken into account to cluster the measured points vertically, to remove the process and sensor noise. The density-based clustering method reduced the noises and allowed detection of the high and low frequency contents of the road.

An adaptive algorithm for generating 3D point clouds of the human body based on 4D millimeter-wave radar.

The traditional algorithms for generating 3D human point clouds often face challenges in dealing with issues such as phantom targets and target classification caused by electromagnetic multipath effects, resulting in a lack of accuracy in the generated point clouds and requiring manual labeling of the position of the human body. To address these problems, this paper proposes an adaptive method for generating 3D human point clouds based on 4D millimeter-wave radar (Self-Adaptive mPoint, SA-mPoint). This method estimates the rough human point cloud by considering micro-motion and respiration characteristics while combining the echo dynamic with static information. Furthermore, it enhances the density of point cloud generation. It reduces interference from multipath noise through multi-frame dynamic fusion and an adaptive density-based clustering algorithm based on the center points of humans. The effectiveness of the SA-mPoint algorithm is verified through experiments conducted using the TI Millimeter Wave Cascade Imaging Radar Radio Frequency Evaluation Module 77G 4D cascade radar to collect challenging raw data consisting of single-target and multi-target human poses in an open classroom setting. Experimental results demonstrate that the proposed algorithm achieves an average accuracy rate of 97.94% for generating point clouds. Compared to the popular TI-mPoint algorithm, it generates a higher number of point clouds on average (increased by 87.94%), improves the average accuracy rate for generating point clouds (increased by 78.3%), and reduces the running time on average (reduced by 11.41%). This approach exhibits high practicality and promising application prospects.

Dynamic Optimization Method of Knowledge Graph Entity Relations for Smart Maintenance of Cantilever Roadheaders

The fault maintenance scenario in coal-mine equipment intelligence is composed of videos, images, signals, and repair process records. Text data are not the primary data that reflect the fault phenomenon, but rather the secondary processing based on operation experience. Focusing on the difficulty of extracting fault knowledge from the limited textual maintenance process records, a forward static full-connected topology network modeling method based on domain knowledge from four dimensions of physical structure, internal association, condition monitoring, and fault maintenance, is proposed to increase the efficiency of constructing a fault-maintenance knowledge graph. Accurately identifying the intrinsic correlation between the equipment anomalies and the faults’ causes through only domain knowledge and loosely coupled data is difficult. Based on the static full-connected knowledge graph of the cantilever roadheader, the information entropy and density-based DBSCAN clustering algorithm is used to process and analyze many condition-monitoring historical datasets to optimize the entity relationships between the fault phenomena and causes. The improved DBSCAN algorithm consists of three stages: firstly, extracting entity data related to fault information from the static fully connected graph; secondly, calculating the information entropy based on the real dataset describing the fault information and the historical operating condition, respectively; and thirdly, comparing the entropy values of the entities and analyzing the intrinsic relationship between the fault phenomenon, the operating condition data, and the fault causes. Based on the static full-connected topology storage in the Neo4j database, the information entropy and density-based DBSCAN algorithm is computed by using Python to identify the relationship weights and dynamically display optimized knowledge graph topology. Finally, an example of EBZ200-type cantilever roadheader for smart maintenance is studied to analyze and evaluate the forward and four-mainlines knowledge graph modeling method and the dynamic entity relations optimization method for static full-connected knowledge graph.

Comparison of two novel methods for counting wheat ears in the field with terrestrial LiDAR

BackgroundThe metrics for assessing the yield of crops in the field include the number of ears per unit area, the grain number per ear, and the thousand-grain weight. Typically, the ear number per unit area contributes the most to the yield. However, calculation of the ear number tends to rely on traditional manual counting, which is inefficient, labour intensive, inaccurate, and lacking in objectivity. In this study, two novel extraction algorithms for the estimation of the wheat ear number were developed based on the use of terrestrial laser scanning (TLS) in conjunction with the density-based spatial clustering (DBSC) algorithm based on the normal and the voxel-based regional growth (VBRG) algorithm. The DBSC involves two steps: (1) segmentation of the point clouds using differences in the normal vectors and (2) clustering of the segmented point clouds using a density clustering algorithm to calculate the ear number. The VBRG involves three steps: (1) voxelization of the point clouds, (2) construction of the topological relationships between the voxels as a connected region using the k-dimensional tree, and (3) detection of the wheat ears in the connected areas using a regional growth algorithm.ResultsThe results demonstrated that DBSC and VBRG were promising in estimating the number of ears for different cultivars, planting densities, N fertilization rates, and growth stages of wheat (RMSE = 76 ~ 114 ears/m2, rRMSE = 18.62 ~ 27.96%, r = 0.76 ~ 0.84). Comparing the performance of the two algorithms, the overall accuracy of the DBSC (RMSE = 76 ears/m2, rRMSE = 18.62%, r = 0.84) was better than that of the VBRG (RMSE = 114 ears/m2, rRMSE = 27.96%, r = 0.76). It was found that with the DBSC, the calculation in points as units permitted more detailed information to be retained, and this method was more suitable for estimation of the wheat ear number in the field.ConclusionsThe algorithms adopted in this study provide new approaches for non-destructive measurement and efficient acquisition of the ear number in the assessment of the wheat yield phenotype.

The triple refinement of self-paced learning style for unsupervised cross-domain person re-identification

Unsupervised cross-domain person re-identification based on clustering typically involves generating pseudo-labels for all training data. However, some data are prone to generating incorrect pseudo-labels in the early stages of training, which leads to a decrease in the accuracy of the model. Based on the idea of self-paced learning, from easy to difficult and from a few to many samples, we propose the Triple Refinement of Self-paced Learning style (TRSL) method to solve this problem. Firstly, we use the density-based clustering algorithm DBSCAN to perform the first data refinement on the unlabeled data. Then, in the second refinement step, we calculate the challenge value by exploring the distance distribution between the clustering data and the centroids to refine the data. In the third refinement step, we calculate the credibility value of the data by comparing the k-nearest neighbor ranking lists of the local and global features of the images, use the credibility value to recall the discarded data from the dual refinement process, and perform a third filtering on the retained data. Finally, the processed data is input into a pair of mutually beneficial learning models for training. On top of a strong baseline, our method performs better than the recent state-of-the-art methods in some evaluation metrics.

Non-iterative border-peeling clustering algorithm based on swap strategy

Border-Peeling algorithm is a recently proposed density based clustering algorithm. The method of peeling off border points by continuous iteration and calculating the density influence value by using the Gaussian kernel distance makes the algorithm more complex. At the same time, there is a risk of excessive peeling of small clusters on unbalanced datasets, which leads to a large number of noise misidentification. In order to reduce the time consumption and improve the clustering performance on unbalanced datasets, this paper proposes a non-iterative border-peeling clustering algorithm. First, the potential core points are determined by the centroids of k-nearest neighbor. Secondly, the points with lower local relative density in the core points and the points with higher relative density in the border points are exchanged to complete the distinction between the core points and the border points. Then basic DBSCAN method is used to cluster core points and noise points. Finally, the associations between the border points and the core points are based on the number of reverse nearest neighbors of the border points in the core points. Our method has achieved competitive results on 10 synthetic datasets and 8 UCI real-world datasets.

Extended methods for spatial cell classification with DBSCAN-CellX

Local cell densities and positioning within cellular monolayers and stratified epithelia have important implications for cell interactions and the functionality of various biological processes. To analyze the relationship between cell localization and tissue physiology, density-based clustering algorithms, such as DBSCAN, allow for a detailed characterization of the spatial distribution and positioning of individual cells. However, these methods rely on predefined parameters that influence the outcome of the analysis. With varying cell densities in cell cultures or tissues impacting cell sizes and, thus, cellular proximities, these parameters need to be carefully chosen. In addition, standard DBSCAN approaches generally come short in appropriately identifying individual cell positions. We therefore developed three extensions to the standard DBSCAN-algorithm that provide: (i) an automated parameter identification to reliably identify cell clusters, (ii) an improved identification of cluster edges; and (iii) an improved characterization of the relative positioning of cells within clusters. We apply our novel methods, which are provided as a user-friendly OpenSource-software package (DBSCAN-CellX), to cellular monolayers of different cell lines. Thereby, we show the importance of the developed extensions for the appropriate analysis of cell culture experiments to determine the relationship between cell localization and tissue physiology.

GFDC: A granule fusion density-based clustering with evidential reasoning

Density-based clustering algorithms are known for their ability to detect irregular clusters, but they have limitations when it comes to dealing with clusters of varying densities. In this paper, we propose a new clustering algorithm called granule fusion density-based clustering with evidential reasoning (GFDC). The approach introduces the concept of sparse degree, which measures both the local density and global density of samples. The sparse degree of samples reflects the stability of samples. Moreover, a core-granule is composed of the neighborhood granule of a sample, of which the sparse degree is minimum in its neighborhood. Then, the core-granules are generated based on the sparse degree and are insensitive to clusters with varying densities. The core samples, which consist of samples in core-granules, are used to form initial clusters through fusion strategies. Additionally, an assignment method is developed from Dempster-Shafer theory to assign border samples and identify outliers. The experimental results demonstrate the effectiveness of GFDC on extensive synthetic and real-world datasets.