Similarity Based Clustering Algorithm Research Articles

A Tracklet-before-Clustering Initialization Strategy Based on Hierarchical KLT Tracklet Association for Coherent Motion Filtering Enhancement

Coherent motions depict the individuals’ collective movements in widely existing moving crowds in physical, biological, and other systems. In recent years, similarity-based clustering algorithms, particularly the Coherent Filtering (CF) clustering approach, have accomplished wide-scale popularity and acceptance in the field of coherent motion detection. In this work, a tracklet-before-clustering initialization strategy is introduced to enhance coherent motion detection. Moreover, a Hierarchical Tracklet Association (HTA) algorithm is proposed to address the disconnected KLT tracklets problem of the input motion feature, thereby making proper trajectories repair to optimize the CF performance of the moving crowd clustering. The experimental results showed that the proposed method is effective and capable of extracting significant motion patterns taken from crowd scenes. Quantitative evaluation methods, such as Purity, Normalized Mutual Information Index (NMI), Rand Index (RI), and F-measure (Fm), were conducted on real-world data using a huge number of video clips. This work has established a key, initial step toward achieving rich pattern recognition.

A framework for the clustering and categorization of CISA reports

This paper presents a framework for text clustering and categorisation. The proposed clustering approach is based on a modified existing similarity-based clustering algorithm, which was originally developed for well-structured data. In this study, the clustering algorithm is used to map text documents into clusters, in order to discover groups of topical documents. The clusters produced in this way are also used for the categorisation of new documents that are uploaded to the system. The algorithms are discussed using as an example the analysis of text documents including Industrial Control Systems (ICS) Advisory Reports and Common Vulnerabilities and Exposures (CVE) recommendations, together available and provided by the Cybersecurity and Infrastructure Security Agency (CISA). Experiments are carried out, although the main focus is on the clustering algorithm. Based on the experimental results, it can be concluded that the proposed similarity-based clustering algorithm can be considered as an alternative approach for text clustering.

Principal component analysis based data collection for sustainable internet of things enabled Cyber–Physical Systems

The Internet of Things (IoT) enabled Cyber–Physical System (CPS) is a promising technology applying in smart home, industrial manufacturing, intelligent transportation, etc. The IoT enabled CPS consists of two main components, i.e., IoT devices and cybers, which interact with each other. The IoT devices collect sensory data from physical environments and transmit them to the cybers, and the cybers make decisions to respond to the collected data and issue commands to control the IoT devices. It is generally known that energy is an important but limited resource in IoT devices. Data compression is an efficient way to reduce the energy consumption of data collection in sustainable IoT enabled CPSs, especially the Principal Component Analysis (PCA) based data compression. The trade-off between data compression ratio and data reconstruction error is one of the biggest challenges for PCA based data compression. In this paper, we investigate PCA based data compression to maximize the compression ratio with bounded reconstruction error for data collection in IoT enabled CPSs. Firstly, a similarity based clustering algorithm is proposed to cluster IoT devices in an IoT enabled CPS. Then, a PCA based data compression algorithm is proposed to compress the collected data to the greatest extent in each cluster with a bounded reconstruction error. Extensive simulations are conducted to verify the efficiency and effectiveness of the proposed algorithms.

An Improved Similarity-Based Clustering Algorithm for Multi-Database Mining.

Clustering algorithms for multi-database mining (MDM) rely on computing pairwise similarities between n multiple databases to generate and evaluate candidate clusterings in order to select the ideal partitioning that optimizes a predefined goodness measure. However, when these pairwise similarities are distributed around the mean value, the clustering algorithm becomes indecisive when choosing what database pairs are considered eligible to be grouped together. Consequently, a trivial result is produced by putting all the n databases in one cluster or by returning n singleton clusters. To tackle the latter problem, we propose a learning algorithm to reduce the fuzziness of the similarity matrix by minimizing a weighted binary entropy loss function via gradient descent and back-propagation. As a result, the learned model will improve the certainty of the clustering algorithm by correctly identifying the optimal database clusters. Additionally, in contrast to gradient-based clustering algorithms, which are sensitive to the choice of the learning rate and require more iterations to converge, we propose a learning-rate-free algorithm to assess the candidate clusterings generated on the fly in fewer upper-bounded iterations. To achieve our goal, we use coordinate descent (CD) and back-propagation to search for the optimal clustering of the n multiple database in a way that minimizes a convex clustering quality measure in less than iterations. By using a max-heap data structure within our CD algorithm, we optimally choose the largest weight variable at each iteration i such that taking the partial derivative of with respect to allows us to attain the next steepest descent minimizing without using a learning rate. Through a series of experiments on multiple database samples, we show that our algorithm outperforms the existing clustering algorithms for MDM.

SDN-Based Centralized Channel Assignment Scheme Using Clustering in Dense WLAN Environments

IEEE WLAN 802.11 uses a contention-based medium access control protocol. Adjacent WLAN access points (APs) and stations (STAs) sharing the same channels causes interference, and affect network performance. Existing distributed channel selection schemes are prone to channel oscillation and ripple effect problems. Additionally, centralized schemes may provide optimal or sub-optimal performance but the computation time complexity increases with number of APs. In this paper, we propose a clustering-based centralized channel selection scheme implemented using software-defined networking architecture to overcome these limitations. In the proposed scheme, we first group the APs using a similarity-based clustering algorithm using proximity information. The channel assignment problem is then formulated using an integer linear programming model. Performance evaluation using simulations indicate that the proposed scheme outperforms the common least congested channel selection (LCCS), in terms of throughput and required number of channels switching. Additionally, for N number of APs, the computation complexity is reduced from $$\mathcal {O}\left( 3^{N}\right)$$ to $$\mathcal {O}\left( (N)^2\right)$$ , in comparison to the common centralized channel selection scheme. The test bed results validate simulation results by achieving better data rates for the STAs compared with the LCCS scheme.

Automated Identification of Chemical Series: Classifying like a Medicinal Chemist.

We investigate different automated approaches for the classification of chemical series in early drug discovery, with the aim of closely mimicking human chemical series conception. Chemical series, which are commonly defined by hand-drawn scaffolds, organize datasets in drug discovery projects. Often, they form the basis for further project decisions. To trace and evaluate these decisions in historic and ongoing projects, it is important to know or reconstruct chemical series. There is not a unique correct definition of chemical series, and the human definition certainly involves a subjective bias. Hence, we first develop quality metrics for the chemical series definitions, evaluating the size and specificity of chemical series. These metrics are applied to categorize human series definitions and implemented in automated classification approaches. For the automated classification of chemical series, we test different fragmentation and similarity-based clustering algorithms and apply different approaches to infer series definitions from these clusters or sets of fragments. We benchmark the classification results against human-defined series from 30 internal projects. The best results in reproducing the composition of human-defined series are achieved when applying UPGMA (unweighted pair group method with arithmetic mean) clustering to the project dataset and calculating maximum common substructures of the clusters as series definitions. We evaluate this approach in more detail on a public dataset and assess its robustness by 10-fold cross-validation, each time sampling 40% of the dataset. Through these benchmarking and validation experiments, we show that the proposed automated approach is able to accurately and robustly identify human-defined series, which comply with a certain, predefined level of specificity and size. Suggesting a thoroughly tested algorithm for series classification, as well as quality metrics for series and several benchmarking approaches, this work lays the foundation for further analysis of project decisions, and it offers an enhanced understanding of the properties of human-defined chemical series.

Multifaceted Traumatic Exposure: Simultaneous Direct and Vicarious Trauma Among EMS Personnel

Graph clustering is an important technique in data mining and network analysis, and it is widely used in chemistry, physics, biology, communication, and computer science. Similarities between vertices of a graph are the fundamental conditions for many hierarchical clustering algorithms. In the paper, we propose a new similarity measure based on microblock density, which computes the similarity between a pair of vertices by calculating the densities of their common adjacent microblock. This measure extends the scope and improves the discrimination of traditional measure, thus significantly improving the performance and stability of the similarity-based clustering algorithms. Experiments on synthetic data and real networks show that the density-based similarity approach accurately reflects the local structure of the graph and provides higher accuracy similarities for clustering and community structural detection algorithms than other state-of-the-art methods.

ANTICALIgN: visualizing, editing and analyzing combined nucleotide and amino acid sequence alignments for combinatorial protein engineering.

ANTIC ALIGN: is an interactive software developed to simultaneously visualize, analyze and modify alignments of DNA and/or protein sequences that arise during combinatorial protein engineering, design and selection. ANTIC ALIGN: combines powerful functions known from currently available sequence analysis tools with unique features for protein engineering, in particular the possibility to display and manipulate nucleotide sequences and their translated amino acid sequences at the same time. ANTIC ALIGN: offers both template-based multiple sequence alignment (MSA), using the unmutated protein as reference, and conventional global alignment, to compare sequences that share an evolutionary relationship. The application of similarity-based clustering algorithms facilitates the identification of duplicates or of conserved sequence features among a set of selected clones. Imported nucleotide sequences from DNA sequence analysis are automatically translated into the corresponding amino acid sequences and displayed, offering numerous options for selecting reading frames, highlighting of sequence features and graphical layout of the MSA. The MSA complexity can be reduced by hiding the conserved nucleotide and/or amino acid residues, thus putting emphasis on the relevant mutated positions. ANTIC ALIGN: is also able to handle suppressed stop codons or even to incorporate non-natural amino acids into a coding sequence. We demonstrate crucial functions of ANTIC ALIGN: in an example of Anticalins selected from a lipocalin random library against the fibronectin extradomain B (ED-B), an established marker of tumor vasculature. Apart from engineered protein scaffolds, ANTIC ALIGN: provides a powerful tool in the area of antibody engineering and for directed enzyme evolution.

Estimation of solar radiation using a combination of Hidden Markov Model and generalized Fuzzy model

Estimation of solar radiation is of considerable importance because of the increasing requirement for the design, optimization and performance evaluation of the solar energy systems. This paper presents the development of pattern similarity based clustering algorithm and its application in solar radiation estimation. In the present work continuous density, Hidden Markov Model (HMM) with Pearson R model is utilized for the extraction of shape based clusters from the input meteorological parameters and it is then processed by the Generalized Fuzzy Model (GFM) to accurately estimate the solar radiation. Instead of using distance function as an index of similarity here shape/patterns of the data vectors are used as the similarity index for clustering, which overcomes few of the shortcomings associated with distance based clustering approaches. The estimation method used here exploits the pattern identification prowess of the HMM for cluster selection and generalization and nonlinear modeling capabilities of GFM to predict the solar radiation. The data of solar radiation and various meteorological parameters (sun shine hour, ambient temperature, relative humidity, wind speed and atmospheric pressure) to carry out the present work is taken from the comprehensive weather monitoring station made at Solar Energy Centre, Gurgaon, India. To consider the effect of each meteorological parameter on the estimation of solar radiation the proposed model is applied on 15 different sets comprising of various combinations of input meteorological parameters. The meteorological data of three years from 2009 to 2011 (915days) is used to estimate the solar radiation. Out of these 915days data, the first 750days data is used for the training of the proposed paradigm and rest 165days data is used for validating the model. The results of estimation using all the sets of various combination of meteorological parameter are analyzed and it is found that the sunshine duration is the prime parameter for the estimation of solar radiation. The next important parameter, which influences the estimation of solar radiation, is temperature followed by relative humidity, atmospheric pressure and wind speed. It is interesting to note that worse results are obtained for the sets which are not using sunshine duration as an input. The best performance is achieved by the set which uses all the parameters except the wind speed. The Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE) and correlation co-efficient (R-value) of the proposed paradigm for the best performing combination of meteorological parameter are 7.9124, 3.0083 and 0.9921 respectively which shows that the proposed model results are in good agreement with the actual measured solar radiation.

Smooth Splicing: A Robust SNN-Based Method for Clustering High-Dimensional Data

Sharing nearest neighbor (SNN) is a novel metric measure of similarity, and it can conquer two hardships: the low similarities between samples and the different densities of classes. At present, there are two popular SNN similarity based clustering methods: JP clustering and SNN density based clustering. Their clustering results highly rely on the weighting value of the single edge, and thus they are very vulnerable. Motivated by the idea of smooth splicing in computing geometry, the authors design a novel SNN similarity based clustering algorithm within the structure of graph theory. Since it inherits complementary intensity-smoothness principle, its generalizing ability surpasses those of the previously mentioned two methods. The experiments on text datasets show its effectiveness.

Variance enhanced K-medoid clustering

This paper proposes new variance enhanced clustering methods to improve the popular K-medoid algorithm by adapting variance information in data clustering. Since measuring similarity between data objects is simpler than mapping data objects to data points in feature space, these pairwise similarity based clustering algorithms can greatly reduce the difficulty in developing clustering based pattern recognition applications. A web-based image clustering system has been developed to demonstrate and show the clustering power and significance of the proposed methods. Synthetic numerical data and real-world image collection are applied to evaluate the performance of the proposed methods on the prototype system. As shown as the web-demonstration, the proposed method, variance enhanced K-medoid model, groups similar images in clusters with various variances according to the distribution of image similarity values.

APPLYING ROBUST DIRECTIONAL SIMILARITY BASED CLUSTERING APPROACH RDSC TO CLASSIFICATION OF GENE EXPRESSION DATA

Despite the fact that the classification of gene expression data from a cDNA microarrays has been extensively studied, nowadays a robust clustering method, which can estimate an appropriate number of clusters and be insensitive to its initialization has not yet been developed. In this work, a novel Robust Clustering approach, RDSC, based on the new Directional Similarity measure is presented. This new approach RDSC, which integrates the Directional Similarity based Clustering Algorithm, DSC, with the Agglomerative Hierarchical Clustering Algorithm, AHC, exhibits its robustness to initialization and its capability to determine the appropriate number of clusters reasonably. RDSC has been successfully employed to both artificial and benchmarking gene expression datasets. Our experimental results demonstrate its distinctive superiority over the conventional method Kmeans and the two typical directional clustering algorithms SPKmeans and moVMF.