CiteULike: mshafiei's library [190 articles]

Bipartite graph partitioning and data clustering

Hongyuan Zha — 2006-04-30T15:04:00-00:00

(2001), pp. 25-32.

GaP: a factor model for discrete data

John Canny — 2006-01-12T09:19:38-00:00

(2004), pp. 122-129.

Recommendation on Item Graphs

Fei Wang — 2008-04-09T19:44:53-00:00

(2006), pp. 1119-1123.

Spectral clustering and transductive learning with multiple views

Dengyong Zhou — 2008-04-09T19:37:58-00:00

(2007), pp. 1159-1166.

Numerical Recipes: The Art of Scientific Computing

William Press — 2007-06-13T05:09:04-00:00

(01 August 2007)

Co-authored by four leading scientists from academia and industry, Numerical Recipes Third Edition starts with basic mathematics and computer science and proceeds to complete, working routines. Widely recognized as the most comprehensive, accessible and practical basis for scientific computing, this new edition incorporates more than 400 Numerical Recipes routines, many of them new or upgraded. The executable C++ code, now printed in color for easy reading, adopts an object-oriented style particularly suited to scientific applications. The whole book is presented in the informal, easy-to-read style that made earlier editions so popular. Find more information at <a href = "http://www.nr.com">www.nr.com or <a href = "http://www.cambridge.org/numericalrecipes">www.cambridge.org/numericalrecipes. New key features: <ul> <li>2 new chapters, 25 new sections, 25% longer than Second Edition</li> <li>Thorough upgrades throughout the text</li> <li>Over 100 completely new routines and upgrades of many more.</li> <li>New Classification and Inference chapter, including Gaussian mixture models, HMMs, hierarchical clustering, Support Vector Machines</li><li>New Computational Geometry chapter covers KD trees, quad- and octrees, Delaunay triangulation, and algorithms for lines, polygons, triangles, and spheres</li> <li>New sections include interior point methods for linear programming, Monte Carlo Markov Chains, spectral and pseudospectral methods for PDEs, and many new statistical distributions</li> <li>An expanded treatment of ODEs with completely new routines</li> </ul> Plus comprehensive coverage of <ul> <li>linear algebra, interpolation, special functions, random numbers, nonlinear sets of equations, optimization, eigensystems, Fourier methods and wavelets, statistical tests, ODEs and PDEs, integral equations, and inverse theory</li> </ul> And much, much more! For more information, or to buy the book, visit <a href = "http://www.cambridge.org/numericalrecipes">www.cambridge.org/numericalrecipes. For support, or to subscribe to an online version, please visit <a href = "http://www.nr.com">www.nr.com.

Topic segmentation with shared topic detection and alignment of multiple documents

Bingjun Sun — 2008-03-31T15:45:10-00:00

(2007), pp. 199-206.

Clustering with Soft and Group Constraints

Martin Law — 2008-03-26T14:07:31-00:00

Structural, Syntactic, and Statistical Pattern Recognition (2004), pp. 662-670.

Several clustering algorithms equipped with pairwise hard constraints between data points are known to improve the accuracy of clustering solutions. We develop a new clustering algorithm that extends mixture clustering in the presence of (i) soft constraints, and (ii) group-level constraints. Soft constraints can reflect the uncertainty associated with a priori knowledge about pairs of points that should or should not belong to the same cluster, while group-level constraints can capture larger building blocks of the target partition when afforded by the side information. Assuming that the data points are generated by a mixture of Gaussians, we derive the EM algorithm to estimate the parameters of different clusters. Empirical study demonstrates that the use of soft constraints results in superior data partitions normally unattainable without constraints. Further, the solutions are more robust when the hard constraints may be incorrect.

A Segment-based Approach To Clustering Multi-Topic Documents

Andrea Tagarelli — 2008-03-01T22:10:31-00:00

(2008)

A tale of two citations

Mounir Errami — 2008-01-24T11:44:34-00:00

Nature, Vol. 451, No. 7177. (24 January 2008), pp. 397-399.

Modelling aging characteristics in citation networks

Kamalika Hajra — 2008-03-24T22:33:12-00:00

Physica A: Statistical Mechanics and its Applications, Vol. 368, No. 2. (15 August 2006), pp. 575-582.

Growing network models with preferential attachment dependent on both age and degree are proposed to simulate certain features of citation network noted in [Redner, arXiv: physics/0407137 (2004)]. In this directed network, a new node gets attached to an older node with the probability ~K(k)f(t) where the degree and age of the older node are k and t, respectively. Several functional forms of K(k) and f(t) have been considered. The desirable features of the citation network can be reproduced with K(k)~k-[beta] and f(t)~exp([alpha]t) with [beta]=2.0 and [alpha]=-0.2 and with simple modifications in the growth scheme.

Aging in citation networks

Kamalika Hajra — 2007-02-07T22:04:31-00:00

Physica A: Statistical Mechanics and its Applications, Vol. 346, No. 1-2. (1 February 2005), pp. 44-48.

In many growing networks, the age of the nodes plays an important role in deciding the attachment probability of the incoming nodes. For example, in a citation network, very old papers are seldom cited while recent papers are usually cited with high frequency. We study actual citation networks to find out the distribution T(t) of t, the time interval between the published and the cited paper. For different sets of data we find a universal behaviour: T(t)~t-0.9 for t[less-than-or-equals, slant]tc and T(t)~t-2 for t>tc where tc~O(10).

The Matrix Stick-Breaking Process: Flexible Bayes Meta-Analysis

David Dunson — 2008-03-21T22:39:53-00:00

pp. 317-327.

In analyzing data from multiple related studies, it often is of interest to borrow information across studies and to cluster similar studies. Although parametric hierarchical models are commonly used, of concern is sensitivity to the form chosen for the random-effects distribution. A Dirichlet process (DP) prior can allow the distribution to be unknown, while clustering studies; however, the DP does not allow local clustering of studies with respect to a subset of the coefficients without making independence assumptions. Motivated by this problem, we propose a matrix stick-breaking process (MSBP) as a prior for a matrix of random probability measures. Properties of the MSBP are considered, and methods are developed for posterior computation using Markov chain Monte Carlo. Using the MSBP as a prior for a matrix of study-specific regression coefficients, we demonstrate advantages over parametric modeling in simulated examples. The methods are further illustrated using a multinational uterotrophic bioassay study.

Using unsupervised learning of a finite Dirichlet mixture model to improve pattern recognition applications

Nizar Bouguila — 2008-03-17T15:18:05-00:00

Pattern Recogn. Lett., Vol. 26, No. 12. (September 2005), pp. 1916-1925.

Bayesian Mixed-Membership Models of Complex and Evolving Networks

EM Airoldi — 2008-03-17T01:59:32-00:00

(2006)

A Bayesian Model for Supervised Clustering with the Dirichlet Process Prior

Hal Iii — 2008-03-12T01:37:16-00:00

J. Mach. Learn. Res., Vol. 6 (2005), pp. 1551-1577.

Mixed-membership models of scientific publications

E Erosheva — 2008-03-11T18:08:56-00:00

(2004)

this document-oriented mixed membership model to a subcollection of the PNAS database supplied to the participants in the Arthur M. Sackler Colloquium on "Mapping Knowledge Domains." We focus in our analysis on a high level description of the fields in Biological Sciences in terms of a small number of extreme or basis categories. Griffiths and Steyvers [6] use a related version of the model for abstracts only and attempt a finer level of description

Special Functions for Applied Scientists

AM Mathai — 2008-03-11T16:43:39-00:00

(07 March 2008)

Chapter 1 introduces elementary classical special functions. Gamma, beta, psi, zeta functions, hypergeometric functions and the associated special functions, generalizations to Meijer's G and Fox's H-functions are examined here. Discussion is confined to basic properties and selected applications. Introduction to statistical distribution theory is provided. Some recent extensions of Dirichlet integrals and Dirichlet densities are discussed. A glimpse into multivariable special functions such as Appell's functions and Lauricella functions is part of Chapter 1. Special functions as solutions of differential equations are examined. Chapter 2 is devoted to fractional calculus. Fractional integrals and fractional derivatives are discussed. Their applications to reaction-diffusion problems in physics, input-output analysis, and Mittag-Leffler stochastic processes are developed. Chapter 3 deals with q-hyper-geometric or basic hypergeometric functions. Chapter 4 covers basic hypergeometric functions and Ramanujan's work on elliptic and theta functions. Chapter 5 examines the topic of special functions and Lie groups. Chapters 6 to 9 are devoted to applications of special functions. Applications to stochastic processes, geometric infinite divisibility of random variables, Mittag-Leffler processes, alpha-Laplace processes, density estimation, order statistics and astrophysics problems, are dealt with in Chapters 6 to 9. Chapter 10 is devoted to wavelet analysis. An introduction to wavelet analysis is given. Chapter 11 deals with the Jacobians of matrix transformations. Various types of matrix transformations and the associated Jacobians are provided. Chapter 12 is devoted to the discussion of functions of matrix argument in the real case. Functions of matrix argument and the pathway models along with their applications are discussed.

Concepts of Independence for Proportions with a Generalization of the Dirichlet Distribution

Robert Connor — 2008-03-10T13:55:18-00:00

Journal of the American Statistical Association, Vol. 64, No. 325. (1969), pp. 194-206.

Statistical entity-topic models

David Newman — 2006-08-30T10:42:33-00:00

(2006), pp. 680-686.

Mixed-membership models of scientific publications.

E Erosheva — 2006-06-16T19:49:08-00:00

Proc Natl Acad Sci U S A, Vol. 101 Suppl 1 (6 April 2004), pp. 5220-5227.

PNAS is one of world's most cited multidisciplinary scientific journals. The PNAS official classification structure of subjects is reflected in topic labels submitted by the authors of articles, largely related to traditionally established disciplines. These include broad field classifications into physical sciences, biological sciences, social sciences, and further subtopic classifications within the fields. Focusing on biological sciences, we explore an internal soft-classification structure of articles based only on semantic decompositions of abstracts and bibliographies and compare it with the formal discipline classifications. Our model assumes that there is a fixed number of internal categories, each characterized by multinomial distributions over words (in abstracts) and references (in bibliographies). Soft classification for each article is based on proportions of the article's content coming from each category. We discuss the appropriateness of the model for the PNAS database as well as other features of the data relevant to soft classification.

Topic transition detection using hierarchical hidden Markov and semi-Markov models

Dinh Phung — 2008-03-06T19:27:27-00:00

(2005), pp. 11-20.

Estimating Hidden Semi-Markov Chains From Discrete Sequences

Y Guedon — 2008-03-07T18:38:57-00:00

pp. 604-639.

This article addresses the estimation of hidden semi-Markov chains from nonstationary discrete sequences. Hidden semi-Markov chains are particularly useful to model the succession of homogeneous zones or segments along sequences. A discrete hidden semi-Markov chain is composed of a nonobservable state process, which is a semi-Markov chain, and a discrete output process. Hidden semi-Markov chains generalize hidden Markov chains and enable the modeling of various durational structures. From an algorithmic point of view, a new forward-backward algorithm is proposed whose complexity is similar to that of the Viterbi algorithm in terms of sequence length (quadratic in the worst case in time and linear in space). This opens the way to the maximum likelihood estimation of hidden semi-Markov chains from long sequences. This statistical modeling approach is illustrated by the analysis of branching and flowering patterns in plants.

Fitting Hidden Semi-Markov Models to Breakpoint Rainfall Data

John Sansom — 2008-03-07T18:35:19-00:00

The paper proposes a hidden semi-Markov model for breakpoint rainfall data that consist of both the times at which rain-rate changes and the steady rates between such changes. The model builds on and extends the seminal work of Ferguson (1980) on variable duration models for speech. For the rainfall data the observations are modelled as mixtures of log-normal distributions within unobserved states where the states evolve in time according to a semi-Markov process. For the latter, parametric forms need to be specified for the state transition probabilities and dwell-time distributions. Recursions for constructing the likelihood are developed and the EM algorithm used to fit the parameters of the model. The choice of dwell-time distribution is discussed with a mixture of distributions over disjoint domains providing a flexible alternative. The methods are also extended to deal with censored data. An application of the model to a large-scale bivariate dataset of breakpoint rainfall measurements at Wellington, New Zealand, is discussed.

Adaptive event detection with time-varying poisson processes

Alexander Ihler — 2008-03-07T18:19:21-00:00

(2006), pp. 207-216.

Catching the Drift: Probabilistic Content Models, with Applications to Generation and Summarization

Regina Barzilay — 2007-01-05T02:54:01-00:00

(2004), pp. 113-120.

A fully Bayesian approach to unsupervised part-of-speech tagging

Sharon Goldwater — 2008-03-07T16:02:27-00:00

(June 2007), pp. 744-751.

Making Sense of Sound: Unsupervised Topic Segmentation over Acoustic Input

Igor Malioutov — 2008-03-07T15:59:59-00:00

(June 2007), pp. 504-511.

Finding document topics for improving topic segmentation

Olivier Ferret — 2008-03-07T15:57:05-00:00

(June 2007), pp. 480-487.

Population Monte Carlo

O Cappe — 2004-12-28T17:03:21-00:00

Journal of Computational & Graphical Statistics, Vol. 13, No. 4., 907.

Handbook of Computational Statistics

JE Gentle — 2008-03-07T13:49:50-00:00

(26 August 2004)

The Handbook of Computational Statistics - Concepts and Methods is divided into 4 parts. It begins with an overview of the field of Computational Statistics, how it emerged as a seperate discipline, how it developed along the development of hard- and software, including a discussion of current active research. The second part presents several topics in the supporting field of statistical computing. Emphasis is placed on the need for fast and accurate numerical algorithms, and it discusses some of the basic methodologies for transformation, data base handling and graphics treatment. The third part focuses on statistical methodology. Special attention is given to smoothing, iterative procedures, simulation and visualization of multivariate data. Finally a set of selected applications like Bioinformatics, Medical Imaging, Finance and Network Intrusion Detection highlight the usefulness of computational statistics.

Semi-Latent Dirichlet Allocation: A Hierarchical Model for Human Action Recognition

Yang Wang — 2008-03-06T19:41:35-00:00

(2007)

Topic transition detection using hierarchical hidden Markov and semi-Markov models

Dinh Phung — 2008-03-06T19:32:34-00:00

(2005), pp. 11-20.

Hidden semi-Markov models (HSMMs)

KP Murphy — 2008-03-06T19:17:20-00:00

Informal Notes (2002)

Segmental semi-markov models and applications to sequence analysis

Xianping Ge — 2008-03-06T19:07:29-00:00

(2002)

Segmental Semi-Markov Models for Change-Point Detection with Applications to Semiconductor Manufacturing

X Ge — 2008-03-06T19:04:30-00:00

(2000)

Towards Visual Exploration of Topic Shifts

Kilian Thiel — 2008-03-03T17:52:12-00:00

Systems, Man and Cybernetics, 2007. ISIC. IEEE International Conference on (2007), pp. 522-527.

This paper presents two approaches to visually analyze the topic shift of a pool of documents over a given period of time. The first of the proposed methods is based on a multi-dimensional scaling algorithm, which places vectors representing terms occurring in certain years (period- frequency-vectors) in a spatial, two-dimensional space. This kind of visualization enables the detection of terms occurring in documents, published in particular years, or terms spread over different years. The second method uses a graph based approach. Publishing dates of documents, as well as their terms are represented by the vertices of a graph. Terms related to a specific publishing year are connected to the vertex of the year via an edge. By usage of activation spreading techniques, terms frequently occurring in documents published in particular years can be discovered visually. We tested both approaches with 2431 abstracts of papers published in the IEEE Transactions on SMC-A, SMC-B, and SMC-C in the years 1996 to 2006. Our experiments indicate that a number of interesting terms can be nicely separated in clumps according to individual years or periods of time. In addition, one can visualize the emergence of specific terms over certain periods of time and how these and other terms fade away again later.

Model-based overlapping clustering

Arindam Banerjee — 2008-03-02T23:01:46-00:00

(2005), pp. 532-537.

Discovering groups of people in Google news

Dhiraj Joshi — 2008-03-03T23:09:18-00:00

(2006), pp. 55-64.

Spatial Latent Dirichlet Allocation

Xiaogang Wang — 2008-02-15T16:34:45-00:00

(2007)

The author-topic model for authors and documents

Rosen Zvi — 2008-03-03T22:56:22-00:00

Proceedings of the 20th conference on Uncertainty in artificial intelligence (2004), pp. 487-494.

Modeling General and Specific Aspects of Documents with a Probabilistic Topic Model

Chaitanya Chemudugunta — 2008-03-03T22:38:06-00:00

(2007), pp. 241-248.

Nested sampling for Potts models

Iain Murray — 2008-03-03T22:15:17-00:00

(2006), pp. 947-954.

Infinite latent feature models and the Indian buffet process

Tom Griffiths — 2008-03-03T22:12:08-00:00

(2006), pp. 475-482.

Bayesian Sets

Zoubin Ghahramani — 2008-03-03T22:10:19-00:00

(2006), pp. 435-442.

Probability and Random Processes

Geoffrey Grimmett — 2005-04-13T15:56:29-00:00

(01 August 2001)

HM-BiTAM: Bilingual Topic Exploration, Word Alignment, and Translation

Bing Zhao — 2008-03-03T14:33:11-00:00

(2007)

An introduction to MCMC for machine learning

C Andrieu — 2006-08-24T15:04:55-00:00

(2003)

This purpose of this introductory paper is threefold. First, it introduces the Monte Carlo method with emphasis on probabilistic machine learning. Second, it reviews the main building blocks of modern Markov chain Monte Carlo simulation, thereby providing and introduction to the remaining papers of this special issue. Lastly, it discusses new interesting research horizons.

Variational MCMC

Nando de Freitas — 2008-03-02T03:38:53-00:00

(2001), pp. 120-127.

Bayesian Data Analysis, Second Edition

Andrew Gelman — 2005-02-27T14:40:12-00:00

(29 July 2003)

Incorporating new and updated information, this second edition of THE bestselling text in Bayesian data analysis continues to emphasize practice over theory, describing how to conceptualize, perform, and critique statistical analyses from a Bayesian perspective. Its world-class authors provide guidance on all aspects of Bayesian data analysis and include examples of real statistical analyses, based on their own research, that demonstrate how to solve complicated problems. Changes in the new edition include: ·Stronger focus on MCMC·Revision of the computational advice in Part III·New chapters on nonlinear models and decision analysis·Several additional applied examples from the authors' recent research·Additional chapters on current models for Bayesian data analysis such as nonlinear models, generalized linear mixed models, and more·Reorganization of chapters 6 and 7 on model checking and data collectionBayesian computation is currently at a stage where there are many reasonable ways to compute any given posterior distribution. However, the best approach is not always clear ahead of time. Reflecting this, the new edition offers a more pluralistic presentation, giving advice on performing computations from many perspectives while making clear the importance of being aware that there are different ways to implement any given iterative simulation computation. The new approach, additional examples, and updated information make Bayesian Data Analysis an excellent introductory text and a reference that working scientists will use throughout their professional life.

An Introduction to Bayesian Analysis: Theory and Methods (Springer Texts in Statistics)

Jayanta Ghosh — 2008-02-29T00:32:04-00:00

(27 July 2006)

This is a graduate-level textbook on Bayesian analysis blending modern Bayesian theory, methods, and applications. Starting from basic statistics, undergraduate calculus and linear algebra, ideas of both subjective and objective Bayesian analysis are developed to a level where real-life data can be analyzed using the current techniques of statistical computing. Advances in both low-dimensional and high-dimensional problems are covered, as well as important topics such as empirical Bayes and hierarchical Bayes methods and Markov chain Monte Carlo (MCMC) techniques. Many topics are at the cutting edge of statistical research. Solutions to common inference problems appear throughout the text along with discussion of what prior to choose. There is a discussion of elicitation of a subjective prior as well as the motivation, applicability, and limitations of objective priors. By way of important applications the book presents microarrays, nonparametric regression via wavelets as well as DMA mixtures of normals, and spatial analysis with illustrations using simulated and real data. Theoretical topics at the cutting edge include high-dimensional model selection and Intrinsic Bayes Factors, which the authors have successfully applied to geological mapping. The style is informal but clear. Asymptotics is used to supplement simulation or understand some aspects of the posterior.