Flat Clustering Research Articles

Fuzzy Membership Partition Based Effective Hierarchical Agglomerative Flat Clustering Method for High Dimensional Data

Fuzzy Membership Partition Based Effective Hierarchical Agglomerative Flat Clustering Method for High Dimensional Data

Development of an Internet-Based Information/DataResource Platform during the COVID-19 Pandemic Using Hierarchical Clustering

There is heavy usage of Internet-based information resource platformsdue to the COVID-19 pandemic. If it can present the search results as a list of groupsof related terms and context, that would be ideal. An additional enhancement would be to display information in the search results in a classified manner. Therefore, the research is proposing an effective and efficient information retrieval system. The methodology adopted follows the conventional Information Retrieval (IR) system design except that Cosine similarity and hierarchical clustering were introduced for search result ranking. The Average Precision of the proposed system was compared with that of flat clustering such as K-means. The result of the evaluation shows that hierarchical clustering improved in precision, being10% better than flat clustering result retrieval result. This led to greater usability, userexperience,and efficiency of the system

Development of an Internet-Based Information/DataResource Platform during the COVID-19 Pandemic Using Hierarchical Clustering

there is heavy usage of Internet-based information resource platformsdue to the COVID-19 pandemic. If it can present the search results as a list of groupsof related terms and context, that would be ideal. An additional enhancement would be to display information in the search results in a classified manner. Therefore, the research is proposing an effective and efficient information retrieval system. The methodology adopted follows the conventional Information Retrieval (IR) system design except that Cosine similarity and hierarchical clustering were introduced for search result ranking. The Average Precision of the proposed system was compared with that of flat clustering such as K-means. The result of the evaluation shows that hierarchical clustering improved in precision, being10% better than flat clustering result retrieval result. This led to greater usability, userexperience,and efficiency of the system.

A Multi‐Granularity Density Peak Clustering Algorithm Based on Variational Mode Decomposition

Density peak clustering (DPC) can identify cluster centers quickly, without any prior knowledge. It is supposed that the cluster centers have a high density and large distance. However, some real datasets have a hierarchical structure, which will result in local cluster centers having a high density but a smaller distance. DPC is a flat clustering algorithm that searches for cluster centers globally, without considering local differences. To address this issue, a Multi-granularity DPC (MG-DPC) algorithm based on Variational mode decomposition (VMD) is proposed. MG-DPC can find global cluster centers in the coarse-grained space, as well as local cluster centers in the fine-grained space. In addition, the density is difficult to calculate when the dataset has a high dimension. Neighborhood preserving embedding (NPE) algorithm can maintain the neighborhood relationship between samples while reducing the dimensionality. Moreover, DPC requires human experience in selecting cluster centers. This paper proposes a method for automatically selecting cluster centers based on Chebyshev’s inequality. MG-DPC is implemented on the dataset of load-data to realize load classification. The clustering performance is evaluated using five validity indices compared with four typical clustering methods. The experimental results demonstrate that MG-DPC outperforms other comparison methods.

Cell type hierarchy reconstruction via reconciliation of multi-resolution cluster tree.

A wealth of clustering algorithms are available for single-cell RNA sequencing (scRNA-seq) data to enable the identification of functionally distinct subpopulations that each possess a different pattern of gene expression activity. Implementation of these methods requires a choice of resolution parameter to determine the number of clusters, and critical judgment from the researchers is required to determine the desired resolution. This supervised process takes significant time and effort. Moreover, it can be difficult to compare and characterize the evolution of cell clusters from results obtained at one single resolution. To overcome these challenges, we built Multi-resolution Reconciled Tree (MRtree), a highly flexible tree-construction algorithm that generates a cluster hierarchy from flat clustering results attained for a range of resolutions. Because MRtree can be coupled with most scRNA-seq clustering algorithms, it inherits the robustness and versatility of a flat clustering approach, while maintaining the hierarchical structure of cells. The constructed trees from multiple scRNA-seq datasets effectively reflect the extent of transcriptional distinctions among cell groups and align well with levels of functional specializations among cells. Importantly, application to fetal brain cells identified subtypes of cells determined mainly by maturation states, spatial location and terminal specification.

A parameter-less algorithm for tensor co-clustering

The majority of the data produced by human activities and modern cyber-physical systems involve complex relations among their features. Such relations can be often represented by means of tensors, which can be viewed as generalization of matrices and, as such, can be analyzed by using higher-order extensions of existing machine learning methods, such as clustering and co-clustering. Tensor co-clustering, in particular, has been proven useful in many applications, due to its ability of coping with n-modal data and sparsity. However, setting up a co-clustering algorithm properly requires the specification of the desired number of clusters for each mode as input parameters. This choice is already difficult in relatively easy settings, like flat clustering on data matrices, but on tensors it could be even more frustrating. To face this issue, we propose a new tensor co-clustering algorithm that does not require the number of desired co-clusters as input, as it optimizes an objective function based on a measure of association across discrete random variables (called Goodman and Kruskal’s tau) that is not affected by their cardinality. We introduce different optimization schemes and show their theoretical and empirical convergence properties. Additionally, we show the effectiveness of our algorithm on both synthetic and real-world datasets, also in comparison with state-of-the-art co-clustering methods based on tensor factorization and latent block models.

Hierarchical Clustering: A Survey

There is a need to scrutinise and retrieve information from data in today's world. Clustering is an analytical technique which involves dividing data into groups of similar objects. Every group is called a cluster, and it is formed from objects that have affinities within the cluster but are significantly different to objects in other groups. The aim of this paper is to look at and compare two different types of hierarchical clustering algorithms. Partition and hierarchical clustering are the two main types of clustering techniques. Hierarchical clustering algorithm is one of the algorithms discussed here. The aforementioned algorithms are described and analysed in terms of factors such as dataset size, data set type, number of clusters formed, consistency, accuracy, and efficiency. Hierarchical clustering is a cluster analysis technique that aims to create a hierarchy of clusters. A hierarchical clustering method is a set of simple (flat) clustering methods arranged in a tree structure. These methods create clusters by recursively partitioning the entities in a top-down or bottom-up manner. We examine and compare hierarchical clustering algorithms in this paper. The intent of discussing the various implementations of hierarchical clustering algorithms is to assist new researchers and beginners to understand how they function, so they can come up with new approaches and innovations for improvement.

Multiple Flat Projections for Cross-Manifold Clustering.

Cross-manifold clustering is an extreme challenge learning problem. Since the low-density hypothesis is not satisfied in cross-manifold problems, many traditional clustering methods failed to discover the cross-manifold structures. In this article, we propose multiple flat projections clustering (MFPC) for cross-manifold clustering. In our MFPC, the given samples are projected into multiple localized flats to discover the global structures of implicit manifolds. Thus, the intersected clusters are distinguished in various projection flats. In MFPC, a series of nonconvex matrix optimization problems is solved by a proposed recursive algorithm. Furthermore, a nonlinear version of MFPC is extended via kernel tricks to deal with a more complex cross-manifold learning situation. The synthetic tests show that our MFPC works on the cross-manifold structures well. Moreover, experimental results on the benchmark datasets and object tracking videos show excellent performance of our MFPC compared with some state-of-the-art manifold clustering methods.

DeepECT: The Deep Embedded Cluster Tree

The idea of combining the high representational power of deep learning techniques with clustering methods has gained much attention in recent years. Optimizing a clustering objective and the dataset representation simultaneously has been shown to be advantageous over separately optimizing them. So far, however, all proposed methods have been using a flat clustering strategy, with the actual number of clusters known a priori. In this paper, we propose the Deep Embedded Cluster Tree (DeepECT), the first divisive hierarchical embedded clustering method. The cluster tree does not need to know the actual number of clusters during optimization. Instead, the level of detail to be analyzed can be chosen afterward and for each sub-tree separately. An optional data-augmentation-based extension allows DeepECT to ignore prior-known invariances of the dataset, such as affine transformations in image data. We evaluate and show the advantages of DeepECT in extensive experiments.

Hierarchically Clustered Representation Learning

The joint optimization of representation learning and clustering in the embedding space has experienced a breakthrough in recent years. In spite of the advance, clustering with representation learning has been limited to flat-level categories, which often involves cohesive clustering with a focus on instance relations. To overcome the limitations of flat clustering, we introduce hierarchically-clustered representation learning (HCRL), which simultaneously optimizes representation learning and hierarchical clustering in the embedding space. Compared with a few prior works, HCRL firstly attempts to consider a generation of deep embeddings from every component of the hierarchy, not just leaf components. In addition to obtaining hierarchically clustered embeddings, we can reconstruct data by the various abstraction levels, infer the intrinsic hierarchical structure, and learn the level-proportion features. We conducted evaluations with image and text domains, and our quantitative analyses showed competent likelihoods and the best accuracies compared with the baselines.

A novel scale-invariant, dynamic method for hierarchical clustering of data affected by measurement uncertainty

An enhanced technique for hierarchical agglomerative clustering is presented. Classical clusterings suffer from non-uniqueness, resulting from the adopted scaling of data and from the arbitrary choice of the function to measure the proximity between elements. Moreover, most classical methods cannot account for the effect of measurement uncertainty on initial data, when present. To overcome these limitations, the definition of a weighted, asymmetric function is introduced to quantify the proximity between any two elements. The data weighting depends dynamically on the degree of advancement of the clustering procedure. The novel proximity measure is derived from a geometric approach to the clustering, and it allows to both disengage the result from the data scaling, and to indicate the robustness of a clustering against the measurement uncertainty of initial data. The method applies to both flat and hierarchical clustering, maintaining the computational cost of the classical methods.

Efficient similarity-based data clustering by optimal object to cluster reallocation.

We present an iterative flat hard clustering algorithm designed to operate on arbitrary similarity matrices, with the only constraint that these matrices be symmetrical. Although functionally very close to kernel k-means, our proposal performs a maximization of average intra-class similarity, instead of a squared distance minimization, in order to remain closer to the semantics of similarities. We show that this approach permits the relaxing of some conditions on usable affinity matrices like semi-positiveness, as well as opening possibilities for computational optimization required for large datasets. Systematic evaluation on a variety of data sets shows that compared with kernel k-means and the spectral clustering methods, the proposed approach gives equivalent or better performance, while running much faster. Most notably, it significantly reduces memory access, which makes it a good choice for large data collections. Material enabling the reproducibility of the results is made available online.

“Anti-Bayesian” flat and hierarchical clustering using symmetric quantiloids

A Pattern Recognition (PR) system that does not involve labelled samples requires the clustering of the samples into their respective classes before the training and testing can be achieved. All of the reported clustering algorithms (except the one reported in Hammer et al., 2015) operate on Bayesian principles, which is understandable because these principles constitute the basis of optimal PR. Recently, Oommen and his co-authors have proposed a novel, counter-intuitive and pioneering PR scheme that is radically opposed to the Bayesian principle. The rational for this paradigm, referred to as the “Anti-Bayesian” (AB) paradigm, involves classification based on the non-central quantiles of the distributions. This paper, extends the results of Hammer et al. (2015) in many directions. Firstly, we generalize our previous AB clustering Hammer et al. (2015) to handle arbitrary d-dimensional spaces using so-called “quantiloids”. Secondly, we extend the AB paradigm to consider how the clustering can be achieved in hierarchical ways, where we analyze both the Top-Down and the Bottom-Up clustering options. Extensive experimentation, on artificial and on challenging real-life data, demonstrates that our clustering achieves results competitive to the state-of-the-art flat, Top-Down and Bottom-Up clustering approaches, demonstrating the power of the AB paradigm.

Stock Market Prediction using Hierarchical Agglomerative and K-Means Clustering Algorithm

Objectives: The stock market performance has more impact on national economy. The purpose of this work is to generate a portfolio to reduce the uncertainty of stock in short term basis. Methods: Hierarchical clustering is more efficient while non-determinism is of concern when compared with flat clustering. Hierarchical agglomerative Clustering is used, which results in more informative structure than flat clustering on unstructured data. Single-link clustering is taken into account as it does not pays more attention to outliers and amalgamation criterion is local than complete-link clustering and results in intuitive cluster structure. Dendrogram is used to represent the progressive formation of clusters in HAC. Findings: Flat clustering K-means algorithm is used to combine the clusters generated by Hierarchical agglomerative clustering (HAC). As the number of samples has been reduced, iterative use of k-means will choose better centroid. Applications: The final list of the recommended stocks is then showcased to the investor on short term basis. The baseline data is downloaded from National Stock Exchange (NSE).

DenPEHC: Density peak based efficient hierarchical clustering

Existing hierarchical clustering algorithms involve a flat clustering component and an additional agglomerative or divisive procedure. This paper presents a density peak based hierarchical clustering method (DenPEHC), which directly generates clusters on each possible clustering layer, and introduces a grid granulation framework to enable DenPEHC to cluster large-scale and high-dimensional (LSHD) datasets. This study consists of three parts: (1) utilizing the distribution of the parameter γ, which is defined as the product of the local density ρ and the minimal distance to data points with higher density δ in “clustering by fast search and find of density peaks” (DPClust), and a linear fitting approach to select clustering centers with the clustering hierarchy decided by finding the “stairs” in the γ curve; (2) analyzing the leading tree (in which each node except the root is led by its parent to join the same cluster) as an intermediate result of DPClust, and constructing the clustering hierarchy efficiently based on the tree; and (3) designing a framework to enable DenPEHC to cluster LSHD datasets when a large number of attributes can be grouped by their semantics. The proposed method builds the clustering hierarchy by simply disconnecting the center points from their parents with a linear computational complexity O(m), where m is the number of clusters. Experiments on synthetic and real datasets show that the proposed method has promising efficiency, accuracy and robustness compared to state-of-the-art methods.

Clustering in non-parametric multivariate analyses

Non-parametric multivariate analyses of complex ecological datasets are widely used. Following appropriate pre-treatment of the data inter-sample resemblances are calculated using appropriate measures. Ordination and clustering derived from these resemblances are used to visualise relationships among samples (or variables). Hierarchical agglomerative clustering with group-average (UPGMA) linkage is often the clustering method chosen. Using an example dataset of zooplankton densities from the Bristol Channel and Severn Estuary, UK, a range of existing and new clustering methods are applied and the results compared. Although the examples focus on analysis of samples, the methods may also be applied to species analysis. Dendrograms derived by hierarchical clustering are compared using cophenetic correlations, which are also used to determine optimum β in flexible beta clustering. A plot of cophenetic correlation against original dissimilarities reveals that a tree may be a poor representation of the full multivariate information. UNCTREE is an unconstrained binary divisive clustering algorithm in which values of the ANOSIM R statistic are used to determine (binary) splits in the data, to form a dendrogram. A form of flat clustering, k-R clustering, uses a combination of ANOSIM R and Similarity Profiles (SIMPROF) analyses to determine the optimum value of k, the number of groups into which samples should be clustered, and the sample membership of the groups. Robust outcomes from the application of such a range of differing techniques to the same resemblance matrix, as here, result in greater confidence in the validity of a clustering approach.

Smartphone image clustering

Every day the use of images from mobile devices as evidence in legal proceedings is more usual and common.Image source acquisition identification is a branch of digital forensic analysis.We use a combination of hierarchical and flat clustering and the use of Sensor Pattern Noise for source identification.We make a series of experiments which emulate similar situations to those that may occur in reality. Every day the use of images from mobile devices as evidence in legal proceedings is more usual and common. Therefore, forensic analysis of mobile device images takes on special importance. This paper explores the branch of forensic analysis which is based on the identification of the source, specifically on the grouping or clustering of images according to their source acquisition. In contrast with other state of the art techniques for source identification, hierarchical clustering does not involve a priori knowledge of the number of images or devices to be identified or training data for a future classification stage. That is, a grouping by classes with all the input images is performed. The proposal is based on the combination of hierarchical and flat clustering and the use of Sensor Pattern Noise (SPN). There has been a series of experiments which emulate similar situations to those that may occur in reality to test the robustness and reliability of the results of the technique. The results are satisfactory in all the experiments, obtaining high rates of success.

A framework for semi-supervised and unsupervised optimal extraction of clusters from hierarchies

We introduce a framework for the optimal extraction of flat clusterings from local cuts through cluster hierarchies. The extraction of a flat clustering from a cluster tree is formulated as an optimization problem and a linear complexity algorithm is presented that provides the globally optimal solution to this problem in semi-supervised as well as in unsupervised scenarios. A collection of experiments is presented involving clustering hierarchies of different natures, a variety of real data sets, and comparisons with specialized methods from the literature.

Disambiguating implicit temporal queries for temporal information retrieval applications

Time is an important dimension of the information retrieval area that can be very useful in helping to meet the users information needs whenever they include temporal intents. However retrieving the information that meets the query demands is not an easy process. The ambiguity of the query is traditionally one of the causes impeding the retrieval of relevant information. This is particularly evident in the case of temporal queries where users tend to be subjective when expressing their intents (e.g., avatar movie instead of avatar movie 2009). Determining the possible times of the query is therefore of the utmost importance when attempting to achieve better disambiguated results and in order to enable new forms of exploring them. In this thesis, we present our contributions to disambiguate implicit temporal queries in realworld environment, i.e. the Web. To understand better this type of queries, three directions may be followed: information extracted from (1) metadata, (2) query logs or (3) document contents. Within the context of this thesis, we will focus on the latter. However, unlike existing approaches we do not resort to a classification methodology. Instead, in our approach, we seek to detect relevant temporal expressions based on corpus statistics and a general similarity measure that makes use of co-occurrences of words and years extracted from the contents of the documents. Moreover, our methodology tends to be mostly language-independent as we do not use any linguistic-based techniques. Instead, we use a rule-based model solution supported by regular expressions. Based on this, we start by performing a comprehensive study of the temporal value of web documents, particularly web snippets, showing that this type of collection is a valuable data source in the process of dating implicit temporal queries. We then develop two methods. A temporal similarity measure to evaluate the correlation between the query and the candidate dates identified, called Generic Temporal Evaluation (GTE) and a threshold-based classifier that selects the most relevant dates while filtering out the non-relevant or incorrect ones, known as GTEClass. Subsequently, we propose two different applications named GTE-Cluster and GTE-Rank. The first one, uses the determined time of the queries to improve search results exploration. For this purpose, we propose a flat temporal clustering model solution where documents are grouped at the year level. GTE-Rank, in turn, uses the same information to temporally re-rank the web search results. We employ a combination approach that considers words and temporal scores, where documents are ranked to reflect the relevance of the snippet for the query, both in the

A Brief Review on Routing in Wireless Sensor Networks

In the last few years the craze of the wireless sensor network (WSN) is increasing and it has become a latest field of research. Protocols used in ad-hoc networks cannot be used in sensor networks because the sensors are tiny in size, have limited battery power, less storage capacity. Hence new protocols are generated for sensor networks. One of the major challenges is to design an energy efficient routing strategy for WSN. In this paper, metric evaluation, designing issues of the sensor routing protocols are discussed. Routing in WSNs can be broadly sub divided in two categories: flat routing and clustering or hierarchical routing. In this paper, several routing schemes are discussed along with some energy aware cluster routing methods. Keywords:Ad-hoc network, metrics, clustering.