Partition Based Clustering Algorithms Research Articles

Data-Driven Approach Toward Airspace Design for Regional Air Mobility Operations in Korea

As the Korean government agencies have recently announced the first official road map for urban air mobility (UAM) with the aim of introducing a new aviation transportation system, many companies in South Korea have initiated projects to design airspace infrastructure for UAM operations at the early stage of development. Given that the agencies tentatively plan to expand UAM to regional air mobility (RAM) operations at the mature stage of development, this research specifically focuses on establishing airspace infrastructure for upcoming RAM operations in South Korea. The proposed methodology leverages three different algorithms: 1) a partitioning-based clustering algorithm for placing vertiport locations, 2) a density-based clustering algorithm for predicting areas of convective weather, and 3) the Latin hypercube sampling-based probabilistic road map (LHS-based PRM) algorithm for generating an adaptive airspace network. The resulting airspace, constructed by the proposed methodology, takes into account airspace restriction areas such as prohibited areas or military operation areas. The main contribution of this research is to employ a data-driven approach using machine learning and LHS-based PRM algorithms to dynamically establish airspace infrastructure to be potentially used for upcoming RAM operations in South Korea.

IMPLEMENTASI K-MEDOIDS DAN MODEL WEIGHTED-LENGTH RECENCY FREQUENCY MONETARY (W-LRFM) UNTUK SEGMENTASI PELANGGAN DILENGKAPI GUI R

The k-medoids algorithm is a partition-based clustering algorithm that groups n objects as much as k clusters. The algorithm uses medoids as the center point (partition) of the cluster. Medoids are actual objects that are randomly selected as the most centered object in a cluster so that the k-medoids algorithm is robust against outliers. Grouping objects in cluster analysis based on similarities between objects. Measurement of similarity between objects can use the euclidean and manhattan distances. The use of distance in cluster analysis can affect cluster results. Validation of cluster results using internal validation, namely the silhouette index. The Weighted-Length Recency Frequency Monetary (W-LRFM) model is a model that applies the relative importance (weight) of the LRFM model according to the importance of each variable in the LRFM model. LRFM model is a model used for customer segmentation based on customer behavior which consists of variables length, recency, frequency, and monetary. The relative importance (weight) of the W-LFRM model uses the Analytics Hierarchical Process (AHP) method. The W-LRFM model is used to calculate the Customer Lifetime Value (CLV) of each cluster. The implementation of k-medoids and the W-LFRM model in this study are used for customer segmentation based on the length, recency frequency, and monetary variable. The formation of these variables is the result of transformation of customer behavior data such as transaction id, date of purchase, and a total amount of 41,073 rows into variable length, recency, frequency, and monetary as much as 5,108 rows. The criteria of the best cluster formed are k = 2 using the manhattan distance with the average of coefficient values = 0.62. The weights on the W-LRFM model produced based on the AHP method are 0.16, 0.29, 0.47, and 0.08 for the variable length, recency, frequency, and monetary. CLV formed from two clusters, namely 0.158 and 0.499. CLV in the second cluster is bigger so that the second cluster becomes the main priority in the marketing strategy. The second cluster has the characteristics 0.29, 0.47, and 0.08 for the variable length, recency, frequency, and monetary. The second cluster has the characteristics means a loyal customer group. The first cluster has characteristics means a potential customer group. This research is assisted by using Graphical User Interface (GUI) R to facilitate analysis

Efficient breast cancer detection using novel intensity partitioning-based clustering algorithm and multi-dimensional LSTM cyclic neural network

Efficient breast cancer detection using novel intensity partitioning-based clustering algorithm and multi-dimensional LSTM cyclic neural network

Adaptive Initialization Method for K-Means Algorithm.

The K-means algorithm is a widely used clustering algorithm that offers simplicity and efficiency. However, the traditional K-means algorithm uses a random method to determine the initial cluster centers, which make clustering results prone to local optima and then result in worse clustering performance. In this research, we propose an adaptive initialization method for the K-means algorithm (AIMK) which can adapt to the various characteristics in different datasets and obtain better clustering performance with stable results. For larger or higher-dimensional datasets, we even leverage random sampling in AIMK (name as AIMK-RS) to reduce the time complexity. 22 real-world datasets were applied for performance comparisons. The experimental results show AIMK and AIMK-RS outperform the current initialization methods and several well-known clustering algorithms. Specifically, AIMK-RS can significantly reduce the time complexity to O (n). Moreover, we exploit AIMK to initialize K-medoids and spectral clustering, and better performance is also explored. The above results demonstrate superior performance and good scalability by AIMK or AIMK-RS. In the future, we would like to apply AIMK to more partition-based clustering algorithms to solve real-life practical problems.

MapReduce Based Crow Search Adopted Partitional Clustering Algorithms For Handling Large Scale Data

Cluster analysis is the prominent data mining technique in knowledge discovery and it discovers the hidden patterns from the data. The K-Means, K-Modes and K-Prototypes are partition based clustering algorithms and these algorithms select the initial centroids randomly. Because of its random selection of initial centroids, these algorithms provide the local optima in solutions. To solve these issues, the strategy of Crow Search algorithm is employed with these algorithms to obtain the global optimum solution. With the advances in information technology, the size of data increased in a drastic manner from terabytes to petabytes. To make proposed algorithms suitable to handle these voluminous data, the phenomena of parallel implementation of these clustering algorithms with Hadoop Mapreduce framework. The proposed algorithms are experimented with large scale data and the results are compared in terms of cluster evaluation measures and computation time with the number of nodes.

MapReduce-Based Crow Search-Adopted Partitional Clustering Algorithms for Handling Large-Scale Data

Cluster analysis is the prominent data mining technique in knowledge discovery and it discovers the hidden patterns from the data. The K-Means, K-Modes and K-Prototypes are partition based clustering algorithms and these algorithms select the initial centroids randomly. Because of its random selection of initial centroids, these algorithms provide the local optima in solutions. To solve these issues, the strategy of Crow Search algorithm is employed with these algorithms to obtain the global optimum solution. With the advances in information technology, the size of data increased in a drastic manner from terabytes to petabytes. To make proposed algorithms suitable to handle these voluminous data, the phenomena of parallel implementation of these clustering algorithms with Hadoop Mapreduce framework. The proposed algorithms are experimented with large scale data and the results are compared in terms of cluster evaluation measures and computation time with the number of nodes.

A mixed data clustering algorithm with noise-filtered distribution centroid and iterative weight adjustment strategy

Clustering is an important technology for data analysis. Cluster analysis for mixed data remains challenging. This paper proposes a mixed data clustering algorithm with noise-filtered distribution centroid and iterative weight adjustment strategy. The proposed algorithm defines noise-filtered distribution centroid for categorical attributes. We combine both mean and noise-filtered distribution centroid to represent the cluster center with mixed attributes, the noise-filtered distribution centroid records the frequency of occurrences for each possible value of the categorical attributes in a cluster more accurately. Furthermore, because the “noise values” are filtered, the measure to calculate the dissimilarity between data objects and cluster centers could be improved. In addition, the algorithm introduces an iterative weight adjustment strategy with combined intra-cluster and inter-cluster information. The unified weight measurement method is used for refining numeric attributes and categorical attributes. Then attributes with higher intra-cluster homogeneity and inter-clusters heterogeneity are considered as attributes with higher priority. They tend to be assigned with relatively heavier weights during clustering. Experimental results on different datasets from the UCI repository show that the MCFCIW algorithm outperforms the existing partition-based clustering algorithm and clustering algorithm based on data conversion for mixed data on both cluster validity indices and convergence speed.

Efficient breast cancer detection using novel intensity partitioning-based clustering algorithm and multi-dimensional LSTM cyclic neural network

Efficient breast cancer detection using novel intensity partitioning-based clustering algorithm and multi-dimensional LSTM cyclic neural network

Hybrid Density- and Partition-Based Clustering Algorithm for Data With Mixed-Type Variables

Clustering is an essential technique for discovering patterns in data. The steady increase in amount and complexity of data over the years led to improvements and development of new clustering algorithms. However, algorithms that can cluster data with mixed variable types (continuous and categorical) remain limited, despite the abundance of data with mixed types particularly in the medical field. Among existing methods for mixed data, some posit unverifiable distributional assumptions or that the contributions of different variable types are not well balanced. We propose a two-step hybrid density- and partition-based algorithm (HyDaP) that can detect clusters after variables selection. The first step involves both density-based and partition-based algorithms to identify the data structure formed by continuous variables and recognize the important variables for clustering; the second step involves partition-based algorithm together with a novel dissimilarity measure we designed for mixed data to obtain clustering results. Simulations across various scenarios and data structures were conducted to examine the performance of the HyDaP algorithm compared to commonly used methods. We also applied the HyDaP algorithm on electronic health records to identify sepsis phenotypes.

Comparison of Clustering K-Means, Fuzzy C-Means, and Linkage for Nasa Active Fire Dataset

One of the causes of forest fires is the lack of speed of handling when a fire occurs. This can be anticipated by determining how many extinguishing units are in the center of the hot spot. To get hotspots, NASA has provided an active fire dataset. The clustering method is used to get the most optimal centroid point. The clustering methods we use are K-Means, Fuzzy C-Means (FCM), and Average Linkage. The reason for using K-means is a simple method and has been applied in various areas. FCM is a partition-based clustering algorithm which is a development of the K-means method. The hierarchical based clustering method is represented by the Average Linkage method. The measurement technique that uses is the sum of the internal distance of each cluster. Elbow evaluation is used to evaluate the optimal cluster. The results obtained after conducting the K-Means trial obtained the best results with a total distance of 145.35 km, and the best clusters from this method were 4 clusters. Meanwhile, the total distance values obtained from the FCM and Linkage methods were 154.13 km and 266.61 km.

Projected fuzzy C-means clustering with locality preservation

Traditional partition-based clustering algorithms, hard or fuzzy version of C-means, could not deal with high-dimensional data sets effectively as redundant features may impact the computation of distances and local spatial structures among patterns are rarely considered. High dimensionality of space gives rise to so-called concentration effect that is detrimental. In this paper, a novel locality preserving based fuzzy C-means (LPFCM) clustering method and its optimization are presented. An orthogonally projected space, which preserves the locality of structural properties, can be generated in LPFCM, thus enhancing the capability of fuzzy C-means (FCM) for handling high-dimensional data. It is the first time to introduce projection techniques to the FCM optimization objective function directly, and the ideas of fuzzy clustering, geometric structure preservation and feature extraction are seamlessly integrated. LPFCM is also regarded as a unified model that combines two separate stages of spectral clustering. Experimental results on some benchmark data sets show the effectiveness of LPFCM in comparison with FCM and some state-of-the-art methods.

A Novel Hybrid PSO-K-Means Clustering Algorithm Using Gaussian Estimation of Distribution Method and Lévy Flight

Clustering is an important data analysis technique, which has been applied to many practical scenarios. However, many partitioning based clustering algorithms are sensitive to the initial state of cluster centroids, may get trapped in a local optimum, and have poor robustness. In recent years, particle swarm optimization (PSO) has been regarded as an effective solution to the problem. However, it has the possibility of converging to a local optimum, especially when solving complex problems. In this paper, we propose a hybrid PSO-K-means algorithm, which uses the Gaussian estimation of distribution method (GEDM) to assist PSO in updating the population information and adopts Levy flight to escape from the local optimum. The proposed algorithm is named a GEDM and Levy flight based PSO-K-means (GLPSOK) clustering algorithm. Firstly, during initialization, a few particles are initialized using the cluster centroids generated by K-means, while other particles are randomly initialized in the search space. Secondly, GEDM and PSO are selected with different probability to update the population information at different optimization stages. Thirdly, Levy flight is adopted to help the search escape from the local optimum. Finally, the greedy strategy is carried out to select the promising particles from the parents and the newly generated candidates. Experimental results on both synthetic data sets and real-world data sets show that the proposed algorithm can produce better clustering results and is more robust than existing classic or state-of-the-art clustering algorithms.

Partitioning and hierarchical based clustering: a comparative empirical assessment on internal and external indices, accuracy, and time

Clustering is an unsupervised data mining technique where exploration is done with little knowledge of data classes. Its aim is to recognize the hidden information from the data for effective decision-making. Though many clustering algorithms has already been implemented till date, still it is an active topic of research for data mining. Researcher’s attempts to explore, compare, evaluate, and improve the different clustering algorithms available, for specialized situation and context. The purpose of all these efforts are to refine and propose improved version of algorithm after statistical evaluation by different metrices. The present research is an attempt to analysis empirically, the partitioning based clustering algorithms and hierarchical based clustering algorithm; by conducting extensive experiments. Both algorithms effectiveness has been measured through external and internal validity indices and Pearson’s correlation distance function using anatomized experiments. The parameters of evaluation that have been taken into consideration; for Internal Indices: Silhouette Index, Davies-Bouldin Validity Index and Calinski-Harabasz index; for external indices: Jaccard index, Rand Index, Entropy and Normalized Mutual Information. The other parameters of evaluation are accuracy and time of execution. Based on the experiments it may be concluded that K-means algorithm produces more promising result than hierarchical algorithm except in accuracy.

Quality Based Analysis of Clustering Algorithms using Diabetes Data for the Prediction of Disease

Clustering is the popular fundamental investigative performance analysis technique commonly used in various applications. The majority of the clustering techniques proved their effectiveness in finding lot of solutions for a variety of datasets. With the aim of test its performance and its clustering qualities are easy to implement by partition based clustering algorithms. The clustering algorithms k-Means and k-Medoids are used to analyze the diabetic datasets and to predict the diseases in this research work. Around 15000 diabetic patient’s consequential final bio-chemistry prescription are taken for the diabetes identification. With number of times executed the run time of the algorithms are compared from the different clusters. Based on their performance the first-rate algorithm in each class was found out.. The best suitable algorithm is suggested for the prediction of diabetes data in this work.

Notice of Retraction Dynamic clustering of data with modified K-prototype algorithm

Notice of Retraction-----------------------------------------------------------------------After careful and considered review of the content of this paper by a duly constituted expert committee, this paper has been found to be in violation of APTIKOM's Publication Principles.We hereby retract the content of this paper. Reasonable effort should be made to remove all past references to this paper.The presenting author of this paper has the option to appeal this decision by contacting ij.aptikom@gmail.com.-----------------------------------------------------------------------Clustering of mixed numerical and categorical data has become a challenge now a days. A number of algorithms dealing with mixed data have been proposed. Speed and simplicity are the two major features that have made the K-Prototype algorithm a famous partition based clustering algorithm. This algorithm has a constraint of providing the value of K initially and sometimes predicting the optimum number of clusters in advance becomes practically impossible. In this paper, a new algorithm based on the K-Prototype algorithm for clustering mixed data with advanced features for automatic generation of appropriate number of clusters is presented.

Pixel Value Graphical Password Scheme: Compatibility of K-means Clustering Algorithm as Pixel Value Password Fault Tolerance Mechanism

In September 2018, the patent for pixel value graphical password scheme was granted in Malaysia. The graphical password scheme was designed to reduce the complexity of previously developed graphical password scheme where a user only requires to load their personal image as password instead of complex graphical challenge during authentication. As the guardian of digital access, Pixel Value Access Control was highly invincible from password pixel forgery attack where a little bit different pixel value derived from loaded image will deny the access. Only the original enrolled image from a registered user can be recognized by Pixel Value Access Control to authenticate the respective username. That fact makes the graphical password scheme is a trusted access control mechanism but, on the other hand, it makes users bound with the only original password pixel image file. Thus, Pixel Value Access Control need to be installed the pixel value fault tolerance mechanism where it could allow users to acquire their password pixel image file from various storage media. The clustering technique was suggested to solve this issue where it allows an altered pixel password being recognized as the same group of the original pixel password. There are number of clustering algorithms developed for various purposed and application of digital image clustering. K-Means algorithm is one the partition-based clustering algorithm that found to be the simplest and fastest clustering algorithm as suggested by many researchers. This paper is mainly to exhibit the selection of K-Means clustering algorithm became the crucial algorithm for Pixel Value Access Control password pixel fault tolerance algorithm. Background of this topic was briefly explained in introduction section, the implementation of K-Means algorithm as Pixel Value Access Control fault tolerance was elaborate in section 2 and followed by validation of the implementation in section 3. At the end of this paper, there is conclusion for this study and followed by list of references.

Initial seed selection for K-modes clustering – A distance and density based approach

Initial seed artefacts play a vital role in proper categorization of the given data set in partitioning based clustering algorithms. Hence, it is important to identify them. We propose a density with distance based method which ensures identification of seed artefacts from different clusters that leads to more accurate clustering results. Our algorithm improves on the search for initial seed artefacts iteratively until the minimum value of the sum of within sum errors, normalized by their data sizes, is ensured. This is because the initial artefacts are selected from different clusters. Here the choice of seed artefacts guarantees a global optimum clustering solution. We have compared our results with random, Wu, Cao and Khan’s methods of initial seed artefact selection, to show the efficacy of our method.

A Trajectory Regression Clustering Technique Combining a Novel Fuzzy C-Means Clustering Algorithm with the Least Squares Method

Rapidly growing GPS (Global Positioning System) trajectories hide much valuable information, such as city road planning, urban travel demand, and population migration. In order to mine the hidden information and to capture better clustering results, a trajectory regression clustering method (an unsupervised trajectory clustering method) is proposed to reduce local information loss of the trajectory and to avoid getting stuck in the local optimum. Using this method, we first define our new concept of trajectory clustering and construct a novel partitioning (angle-based partitioning) method of line segments; second, the Lagrange-based method and Hausdorff-based K-means++ are integrated in fuzzy C-means (FCM) clustering, which are used to maintain the stability and the robustness of the clustering process; finally, least squares regression model is employed to achieve regression clustering of the trajectory. In our experiment, the performance and effectiveness of our method is validated against real-world taxi GPS data. When comparing our clustering algorithm with the partition-based clustering algorithms (K-means, K-median, and FCM), our experimental results demonstrate that the presented method is more effective and generates a more reasonable trajectory.

Energy aware clustered load balancing in cloud computing environment

Cloud is a collection of datacentres with heterogeneous resources, which gives services to the users based on pay-as-you-use model. Even though it has several advantages such as availability, scalability, and reliability, some performance parameters like energy consumption, load balancing, response time, resource allocation time, etc., are not properly fine tuned. This paper proposes an energy aware clustered load balancing system in which, heterogeneous resources are clustered into different groups by using a partitioning-based clustering algorithm. The clustering reduces number of resources needs to be searched and hence minimizes the time required for resource discovery. An energy aware best-fit virtual machine (VM) allocation is used for reducing the power consumption. The process allocations to VMs are done based on best-fit allocation strategy for optimal space utilisation. The results show that proposed method reduces time for resource discovery, resource allocation and response time with power consumption.

Energy aware clustered load balancing in cloud computing environment

Cloud is a collection of datacentres with heterogeneous resources, which gives services to the users based on pay-as-you-use model. Even though it has several advantages such as availability, scalability, and reliability, some performance parameters like energy consumption, load balancing, response time, resource allocation time, etc., are not properly fine tuned. This paper proposes an energy aware clustered load balancing system in which, heterogeneous resources are clustered into different groups by using a partitioning-based clustering algorithm. The clustering reduces number of resources needs to be searched and hence minimizes the time required for resource discovery. An energy aware best-fit virtual machine (VM) allocation is used for reducing the power consumption. The process allocations to VMs are done based on best-fit allocation strategy for optimal space utilisation. The results show that proposed method reduces time for resource discovery, resource allocation and response time with power consumption.