Searcharxiv⌕ Search

SEARCH · Searcharxiv

Results for “stat.OT”

Search indexed arXiv papers on artificial intelligence, large language models, computer vision and robotics. Read source abstracts and follow links to arXiv.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 109 records · Page 6Linked to original sources
arXiv2014

A Unified Community Detection, Visualization and Analysis method

Community detection in social graphs has attracted researchers' interest for a long time. With the widespread of social networks on the Internet it has recently become an important research domain. Most contributions focus upon the definition of algorithms for optimizing the so-called modularity function. In the first place interest was limited to unipartite graph inputs and partitioned community outputs. Recently bipartite graphs, directed graphs and overlapping communities have been investigated. Few contributions embrace at the same time the three types of nodes. In this paper we present a method which unifies commmunity detection for the three types of nodes and at the same time merges partitionned and overlapping communities. Moreover results are visualized in such a way that they can be analyzed and semantically interpreted. For validation we experiment this method on well known simple benchmarks. It is then applied to real data in three cases. In two examples of photos sets with tagged people we reveal social networks. A second type of application is of particularly interest. After applying our method to Human Brain Tractography Data provided by a team of neurologists, we produce clusters of white fibers in accordance with other well known clustering methods. Moreover our approach for visualizing overlapping clusters allows better understanding of the results by the neurologist team. These last results open up the possibility of applying community detection methods in other domains such as data analysis with original enhanced performances.

cs.SI↗
arXiv2013

Propagation of initial errors on the parameters for linear and Gaussian state space models

For linear and Gaussian state space models parametrized by $θ_0 \in Θ\subset \mathbb{R}^r, r \geq 1$ corresponding to the vector of parameters of the model, the Kalman filter gives exactly the solution for the optimal filtering under weak assumptions. This result supposes that $θ_0$ is perfectly known. In most real applications, this assumption is not realistic since $θ_0$ is unknown and has to be estimated. In this paper, we analysis the Kalman filter for a biased estimator of $θ_0$. We show the propagation of this bias on the estimation of the hidden state. We give an expression of this propagation for linear and Gaussian state space models and we extend this result for almost linear models estimated by the Extended Kalman filter. An illustration is given for the autoregressive process with measurement noises widely studied in econometrics to model economic and financial data.

stat.OT↗
arXiv2013

A Peer-based Model of Fat-tailed Outcomes

It is well known that the distribution of returns from various financial instruments are leptokurtic, meaning that the distributions have "fatter tails" than a Normal distribution, and have skew toward zero. This paper presents a graceful micro-level explanation for such fat-tailed outcomes, using agents whose private valuations have Normally-distributed errors, but whose utility function includes a term for the percentage of others who also buy.

q-fin.TR↗
arXiv2013

EM-based Semi-blind Channel Estimation in AF Two-Way Relay Networks

We propose an expectation maximization (EM)-based algorithm for semi-blind channel estimation of reciprocal channels in amplify-and-forward (AF) two-way relay networks (TWRNs). By incorporating both data samples and pilots into the estimation, the proposed algorithm provides substantially higher accuracy than the conventional training-based approach. Furthermore, the proposed algorithm has a linear computational complexity per iteration and converges after a small number of iterations.

cs.IT↗
arXiv2013

Bayesian Modeling and MCMC Computation in Linear Logistic Regression for Presence-only Data

Presence-only data are referred to situations in which, given a censoring mechanism, a binary response can be observed only with respect to on outcome, usually called \textit{presence}. In this work we present a Bayesian approach to the problem of presence-only data based on a two levels scheme. A probability law and a case-control design are combined to handle the double source of uncertainty: one due to the censoring and one due to the sampling. We propose a new formalization for the logistic model with presence-only data that allows further insight into inferential issues related to the model. We concentrate on the case of the linear logistic regression and, in order to make inference on the parameters of interest, we present a Markov Chain Monte Carlo algorithm with data augmentation that does not require the a priori knowledge of the population prevalence. A simulation study concerning 24,000 simulated datasets related to different scenarios is presented comparing our proposal to optimal benchmarks.

stat.CO↗
arXiv2013

Review of: Mixed Effects Models and Extensions in Ecology with R

This is a review of the book "Mixed Effects Models and Extensions in Ecology with R" by Zuur, Ieno, Walker, Saveliev and Smith (2009, Springer). I was asked to review this book for The American Statistician in 2010. After I wrote the review, the invitation was revoked. This is the review.

stat.OT↗
arXiv2013

Evolutionary Model of a Anonymous Consumer Durable Market

An analytic model is presented that considers the evolution of a market of durable goods. The model suggests that after introduction goods spread always according to a Bass diffusion. However, this phase will be followed by a diffusion process for durable consumer goods governed by a variation-selection-reproduction mechanism and the growth dynamics can be described by a replicator equation. Describing the aggregate sales as the sum of first, multiple and replacement purchase the product life cycle can be derived. Replacement purchase causes periodic variations of the sales determined by the finite lifetime of the good (Juglar cycles). The model suggests that both, Bass- and Gompertz diffusion may contribute to the product life cycle of a consumer durable. The theory contains the standard equilibrium view of a market as a special case. It depends on the time scale, whether an equilibrium or evolutionary description is more appropriate. The evolutionary framework is used to derive also the size, growth rate and price distribution of manufacturing business units. It predicts that the size distribution of the business units (products) is lognormal, while the growth rates exhibit a Laplace distribution. Large price deviations from the mean price are also governed by a Laplace distribution (fat tails). These results are in agreement with empirical findings. The explicit comparison of the time evolution of consumer durables with empirical investigations confirms the close relationship between price decline and Gompertz diffusion, while the product life cycle can be described qualitatively for a long time period.

q-fin.GN↗
arXiv2013

Decision Making for Inconsistent Expert Judgments Using Negative Probabilities

In this paper we provide a simple random-variable example of inconsistent information, and analyze it using three different approaches: Bayesian, quantum-like, and negative probabilities. We then show that, at least for this particular example, both the Bayesian and the quantum-like approaches have less normative power than the negative probabilities one.

stat.OT↗
arXiv2013

Is there currently a scientific revolution in scientometrics?

The author of this letter to the editor would like to set forth the argument that scientometrics is currently in a phase in which a taxonomic change, and hence a revolution, is taking place. One of the key terms in scientometrics is scientific impact which nowadays is understood to mean not only the impact on science but the impact on every area of society.

cs.DL↗
arXiv2013

The normalization of citation counts based on classification systems

If we want to assess whether the paper in question has had a particularly high or low citation impact compared to other papers, the standard practice in bibliometrics is to normalize citations in respect of the subject category and publication year. A number of proposals for an improved procedure in the normalization of citation impact have been put forward in recent years. Against the background of these proposals this study describes an ideal solution for the normalization of citation impact: in a first step, the reference set for the publication in question is collated by means of a classification scheme, where every publication is associated with a single principal research field or subfield entry (e. g. via Chemical Abstracts sections) and a publication year. In a second step, percentiles of citation counts are calculated for this set and used to assign the normalized citation impact score to the publications (and also to the publication in question).

cs.DL↗
arXiv2013

On the asymptotics of Ajtai-Komlós-Tusnády statistics

In our days there is a widespread analysis of Wasserstein distances between theoretical and empirical measures. One of the first investigation of the topic is given in the paper written by Ajtai, Komlós and Tusnády in $1984.$ Interestingly, all the neighboring questions posed by that paper were settled already without the original one. In this paper we are going to delineate the limit behavior of the original statistics with the help of computer simulations. At the same time we kept an eye on theoretical grasping of the problem. Based on our computer simulations our opinion is that the limit distribution is Gaussian.

stat.OT↗
arXiv2013

The factor paradox: Common factors can be correlated with the variance not accounted for by the common factors!

The case that the factor model does not account for all the covariances of the observed variables is considered. This is a quite realistic condition because some model error as well as some sampling error should usually occur with empirical data. It is shown that principal components representing covariances not accounted for by the factors of the model can have a non-zero correlation with the common factors of the factor model. Non-zero correlations of components representing variance not accounted for by the factor model with common factors were also found in a simulation study. Based on these results it should be concluded that common factors can be correlated with variance components representing model error as well as sampling error. In consequence, even when researchers decide not to represent some small or trivial variance by means of a common factor, these excluded variances can still be part of the model.

stat.AP↗
arXiv2013

Meeting Student Needs for Multivariate Data Analysis: A Case Study in Teaching a Multivariate Data Analysis Course with No Pre-requisites

Modern students encounter big, messy data sets long before setting foot in our classrooms. Many of our students need to develop skills in exploratory data analysis and multivariate analysis techniques for their jobs after college, but these topics are not covered in introductory statistics courses. This case study describes my experience in designing and teaching a course on multivariate data analysis with no pre-requisites, using real data, active learning, and other activities to help students tackle the material.

stat.OT↗
arXiv2013

Generalized Thompson Sampling for Contextual Bandits

Thompson Sampling, one of the oldest heuristics for solving multi-armed bandits, has recently been shown to demonstrate state-of-the-art performance. The empirical success has led to great interests in theoretical understanding of this heuristic. In this paper, we approach this problem in a way very different from existing efforts. In particular, motivated by the connection between Thompson Sampling and exponentiated updates, we propose a new family of algorithms called Generalized Thompson Sampling in the expert-learning framework, which includes Thompson Sampling as a special case. Similar to most expert-learning algorithms, Generalized Thompson Sampling uses a loss function to adjust the experts' weights. General regret bounds are derived, which are also instantiated to two important loss functions: square loss and logarithmic loss. In contrast to existing bounds, our results apply to quite general contextual bandits. More importantly, they quantify the effect of the "prior" distribution on the regret bounds.

cs.LG↗
arXiv2013

Hypergeometric tail inequalities: ending the insanity

The hypergeometric distribution is briefly and informally surveyed, including popular notation, symmetries, and the tail inequalities $Pr[i \ge E[i]+tn] \le e^{-2t^2n}$ and $Pr[i \le E[i]-tn] \le e^{-2t^2n}$.

math.PR↗
arXiv2013

Filters and smoothers for self-exciting Markov modulated counting processes

We consider a self-exciting counting process, the parameters of which depend on a hidden finite-state Markov chain. We derive the optimal filter and smoother for the hidden chain based on observation of the jump process. This filter is in closed form and is finite dimensional. We demonstrate the performance of this filter both with simulated data, and by analysing the `flash crash' of 6th May 2010 in this framework.

q-fin.CP↗
arXiv2013

An agent-based epidemiological model of incarceration

We build an agent-based model of incarceration based on the SIS model of infectious disease propagation. Our central hypothesis is that the observed racial disparities in incarceration rates between Black and White Americans can be explained as the result of differential sentencing between the two demographic groups. We demonstrate that if incarceration can be spread through a social influence network, then even relatively small differences in sentencing can result in the large disparities in incarceration rates. Controlling for effects of transmissibility, susceptibility, and influence network structure, our model reproduces the observed large disparities in incarceration rates given the differences in sentence lengths for White and Black drug offenders in the United States without extensive parameter tuning. We further establish the suitability of the SIS model as applied to incarceration, as the observed structural patterns of recidivism are an emergent property of the model. In fact, our model shows a remarkably close correspondence with California incarceration data, without requiring any parameter tuning. This work advances efforts to combine the theories and methods of epidemiology and criminology.

stat.AP↗