Model-based clustering techniques have been widely used and have shown promising results in many applications involving complex data. This paper presents a unified framework for probabilistic model-based clustering based on a bipartite graph view of data and models that highlights the commonalities and differences among existing model-based clustering algorithms. In this view, clusters are represented as probabilistic models in a model space that is conceptually separate from the data space. For partitional clustering, the view is conceptually similar to the ExpectationMaximization (EM) algorithm. For hierarchical clustering, the graph-based view helps to visualize critical/important distinctions between similarity-based approaches and model-based approaches.

The Dirichlet compound multinomial (DCM) distribution, also called the multivariate Polya distribution, is a model for text documents that takes into account burstiness: the fact that if a word occurs once in a document, it is likely to occur repeatedly. We derive a new family of distributions that are approximations to DCM distributions and constitute an exponential family, unlike DCM distributions. We use these so-called EDCM distributions to obtain insights into the properties of DCM distributions, and then derive an algorithm for EDCM maximum-likelihood training that is many times faster than the corresponding method for DCM distributions. Next, we investigate expectationmaximization with EDCM components and deterministic annealing as a new clustering algorithm for documents. Experiments show that the new algorithm is competitive with the best methods in the literature, and superior from the point of view of finding models with low perplexity. 1.

Topic modeling techniques have widespread use in text data mining applications. Some applications use batch models, which perform clustering on the document collection in aggregate. In this paper, we analyze and compare the performance of three recently-proposed batch topic models—Latent Dirichlet Allocation (LDA), Dirichlet Compound Multinomial (DCM) mixtures and von-Mises Fisher (vMF) mixture models. In cases where offline clustering on complete document collections is infeasible due to resource and response-rate constraints, online unsupervised clustering methods that process incoming data incrementally are necessary. To this end, we propose online variants of vMF, EDCM and LDA. Experiments on large real-world document collections, in both the offline and online settings, demonstrate that though LDA is a good model for finding word-level topics, vMF finds better document-level topic clusters more efficiently, which is often important in text mining applications. Finally, we propose a practical heuristic for hybrid topic modeling, which learns online topic models on streaming text and intermittently runs batch topic models on aggregated documents offline. Such a hybrid model is useful for several applications (e.g., dynamic topic-based aggregation of user-generated content in social networks) that need a good tradeoff between the performance of batch offline algorithms and efficiency of incremental online algorithms. 1

Resource discovery is an important process for finding suitable nodes that satisfy application requirements in large loosely-coupled distributed systems. Besides inter-node heterogeneity, many of these systems also show a high degree of intra-node dynamism, so that selecting nodes based only on their recently observed resource capacities for scalability reasons can lead to poor deployment decisions resulting in application failures or migration overheads. In this paper, we propose the notion of a resource bundle— a representative resource usage distribution for a group of nodes with similar resource usage patterns—that employs two complementary techniques to overcome the limitations of existing techniques: resource usage histograms to provide statistical guarantees for resource capacities, and clustering-based resource aggregation to achieve scalability. Using trace-driven simulations and data analysis of a month-long PlanetLab trace, we show that resource bundles are able to provide high accuracy for statistical resource discovery (up to 56 % better precision than using only recent values), while achieving high scalability (up to 55% fewer messages than a non-aggregation algorithm). We also show that resource bundles are ideally suited for identifying group-level characteristics such as finding load hot spots and estimating total group capacity (within 8 % of actual values). 1.

Abstract not found

www.library.drexel.edu The following item is made available as a courtesy to scholars by the author(s) and Drexel University Library and may contain materials and content, including computer code and tags, artwork, text, graphics, images, and illustrations (Material) which may be protected by copyright law. Unless otherwise noted, the Material is made available for non profit and educational purposes, such as research, teaching and private study. For these limited purposes, you may reproduce (print, download or make copies) the Material without prior permission. All copies must include any copyright notice originally included with the Material. You must seek permission from the authors or copyright owners for all uses that are not allowed by fair use and other provisions of the U.S. Copyright Law. The responsibility for making an independent legal assessment and securing any necessary permission rests with persons desiring to reproduce or use the Material. Please direct questions to archives@drexel.edu

Abstract. In this paper we introduce a novel document clustering approach that solves some major problems of traditional document clustering approaches. Instead of depending on traditional vector space model, this approach represents a set of documents as bipartite graphs using domain knowledge in ontology. In this representation, the concepts of the documents are classified according to their relationships with documents that are reflected on the bipartite graph. Using the concept groups, documents are clustered based on the concepts ’ contribution to each document. Through the mutual-refinement relationship with concept groups and document groups, the two groups are recursively refined. Our experimental results on MEDLINE articles show that our approach outperforms two leading document clustering algorithms: BiSecting K-means and CLUTO. In addition to its decent performance, our approach provides a meaningful explanation for each document cluster by identifying its most contributing concepts, thus helps users to understand and interpret documents and clustering results. 1

Abstract — One way to handle data mining problems where class prior probabilities and/or misclassification costs between classes are highly unequal is to resample the data until a new, desired class distribution in the training data is achieved. Many resampling techniques have been proposed in the past, and the relationship between resampling and cost-sensitive learning has been well studied. Surprisingly, however, few resampling techniques attempt to create new, artificial data points which generalize the known, labeled data. In this paper, we introduce an easily implementable resampling technique (generative oversampling) which creates new data points by learning from available training data. Empirically, we demonstrate that generative oversampling outperforms other wellknown resampling methods on several datasets in the example domain of text classification. I.

As the number of RSS news feeds continue to increase over the Internet, it becomes necessary to minimize the workload of the user who is otherwise required to scan through huge number of news articles to find related articles of interest, which is a tedious and often an impossible task. In order to solve this problem, we present a novel approach, called InFRSS, which consists of a correlation-based phrase matching (CPM) model and a fuzzy compatibility clustering (FCC) model. CPM can detect RSS news articles containing phrases that are the same as well as semantically alike, and dictate the degrees of similarity of any two articles. FCC identifies and clusters non-redundant, closely related RSS news articles based on their degrees of similarity and a fuzzy compatibility relation. Experimental results show that (i) our CPM model on matching bigrams and trigrams in RSS news articles outperforms other phrase/keyword-matching approaches and (ii) our FCC model generates high quality clusters and outperforms other well-known clustering techniques.

Resource discovery is an important process for finding suitable nodes that satisfy application requirements in large loosely-coupled distributed systems. Besides inter-node heterogeneity, many of these systems also show a high degree of intra-node dynamism, so that selecting nodes based only on their recently observed resource capacities can lead to poor deployment decisions resulting in application failures or migration overheads. However, most existing resource discovery mechanisms rely mainly on recent observations to achieve scalability in large systems. In this paper, we propose the notion of a resource bundle- a representative resource usage distribution for a group of nodes with similar resource usage patterns- that employs two complementary techniques to overcome the limitations of existing techniques: resource usage histograms to provide statistical guarantees for resource capacities, and clustering-based resource aggregation to achieve scalability. Using trace-driven simulations and data analysis of a month-long PlanetLab trace, we show that resource bundles are able to provide high accuracy for statistical resource discovery, while achieving high scalability. We also show that resource bundles are ideally suited for identifying group-level characteristics (e.g. hot spots, total group capacity). To automatically parameterize the bundling algorithm, we present an adaptive algorithm that can detect online fluctuations in resource heterogeneity. Index Terms resource discovery, aggregation, resource management, machine learning