Kevin Drummey

Loyola University Maryland, Mathematics and Statistics, Faculty Member

Followers

Following

Public Views

Universidad de Guanajuato

Johann Brain

Universidad Arturo Prat

anmol raj

Interests

Uploads

Papers by Kevin Drummey

Robust Bayesian estimation in the normal, gamma and binomial probability models: a computational approach

In this dissertation we address the general problem of robust Bayesian estimation of certain one-... more In this dissertation we address the general problem of robust Bayesian estimation of certain one-dimensional parameters in three important statistical scenarios. Assuming the robust Bayesian perspective of Berger (1984), we find robust Bayes estimators of the normal mean (with known variance), the normal variance (with known mean) and the binomial success probability (with fixed number of trials). Our solution to the normal variance problem concurrently applies to the equivalent problem of robust Bayesian estimation of the gamma scale parameter (with known shape parameter). In addition, the likelihood principle, as discussed in Berger and Wolpert (1988), guarantees the suitability of our binomial problem solution to the negative binomial (including geometric) distribution as well. This dissertation admittedly reflects state-of-the-art computer technology, which gives it a rich computational flavor. We develop FORTRAN programs that employ importance sampling techniques of Stewart (19...

Download

Effective Estimation of Posterior Probabilities: Explaining the Accuracy of Randomized Decision Tree Approaches

Fifth IEEE International Conference on Data Mining (ICDM'05)

There has been increasing number of independently proposed randomization methods in different sta... more There has been increasing number of independently proposed randomization methods in different stages of decision tree construction to build multiple trees. Randomized decision tree methods have been reported to be significantly more accurate than widely-accepted single decision trees, although the training procedure of some methods incorporates a surprisingly random factor and therefore opposes the generally accepted idea of employing gain functions to choose optimum features at each node and compute a single tree that fits the data. One important question that is not well understood yet is the reason behind the high accuracy. We provide an insight based on posterior probability estimations. We first establish the relationship between effective posterior probability estimation and effective loss reduction. We argue that randomized decision tree methods effectively approximate the true probability distribution using the decision tree hypothesis space. We conduct experiments using both synthetic and real-world datasets under both 0-1 and cost-sensitive loss functions.

Download

Agent learning in the multi-agent contracting system [MACS]

Decision Support Systems, 2008

This paper presents a Bayesian learning approach for a multi-agent system, called multi-agent con... more

A comparison of rankings produced by summarization evaluation measures

NAACL-ANLP 2000 Workshop on Automatic summarization -, 2000

evaluation measures produce a ranking of all possible extract summaries of a document., Recall-ba... more evaluation measures produce a ranking of all possible extract summaries of a document., Recall-based evaluation measures, which depend on costly human-generated ground truth summaries, produce uncorrelated rankings when ground truth is varied. This paper proposes using sentence-rankbased and content-based measures for evaluating extract summaries, and compares these with recallbased evaluation measures. Content-based measures increase the correlation of rankings induced by synonymous ground truths, and exhibit other desirable properties.

Download

Robust Bayesian estimation in the normal, gamma and binomial probability models: a computational approach

Download

Effective Estimation of Posterior Probabilities: Explaining the Accuracy of Randomized Decision Tree Approaches

Fifth IEEE International Conference on Data Mining (ICDM'05)

Download

Agent learning in the multi-agent contracting system [MACS]

Decision Support Systems, 2008

This paper presents a Bayesian learning approach for a multi-agent system, called multi-agent con... more

A comparison of rankings produced by summarization evaluation measures

NAACL-ANLP 2000 Workshop on Automatic summarization -, 2000

Download

Kevin Drummey

Related Authors

Uploads

Papers by Kevin Drummey

Log In