Results 1  10
of
10
Some inconsistencies and misidentified modelling assumptions in probabilistic information retrieval
 A CM Transactions on Information Systems
, 1995
"... Research in the probabilistic theory of information retrieval involves the construction of mathematical models based on statistical assumptions. One of the hazards inherent in this kind of theory construction is that the assumptions laid down maybe inconsmtent in unanticipated ways with the data to ..."
Abstract

Cited by 28 (0 self)
 Add to MetaCart
Research in the probabilistic theory of information retrieval involves the construction of mathematical models based on statistical assumptions. One of the hazards inherent in this kind of theory construction is that the assumptions laid down maybe inconsmtent in unanticipated ways with the data to which they are applied. Another hazard is that the stated assumptions may not be those on which the derived modeling equations or resulting experiments are actually based. Both kinds of mistakes have been made m past research on probabihstic reformation retrieval. One consequence of these errors is that the statistical character of certain probabilistic IR models, including the socalled Binary Independence model, has been seriously misapprehended Categories and Subject Descriptors: H. 1.2 [Models and Principles]: User/Machine Systems;
Term Dependence: Truncating the Bahadur Lazarsfeld Expansion
 Information Processing and Management
, 1994
"... The performance of probabilistic information retrieval systems is studied where differing statistical dependence assumptions are used when estimating the probabilities inherent in the retrieval model. Experimental results using the Bahadur Lazarsfeld expansion suggest that the greatest degree of ..."
Abstract

Cited by 17 (7 self)
 Add to MetaCart
The performance of probabilistic information retrieval systems is studied where differing statistical dependence assumptions are used when estimating the probabilities inherent in the retrieval model. Experimental results using the Bahadur Lazarsfeld expansion suggest that the greatest degree of performance increase is achieved by incorporating term dependence information in estimating . It is suggested that incorporating dependence in to degree 3 be used; incorporating more dependence information results in relatively little increase in performance. Experiments examine the span of dependence in natural language text, the window of terms in which dependencies are computed and their effect on information retrieval performance. Results provide additional support for the notion of a window of to terms in width; terms in this window may be most useful when computing dependence. 2 1 Introduction Those who study information retrieval often assume that the features or terms use...
The Maximum Entropy Approach and Probabilistic IR Models
 ACM TRANSACTIONS ON INFORMATION SYSTEMS
, 1998
"... The Principle of Maximum Entropy is discussed and two classic probabilistic models of information retrieval, the Binary Independence Model of Robertson and Sparck Jones and the Combination Match Model of Croft and Harper are derived using the maximum entropy approach. The assumptions on which the cl ..."
Abstract

Cited by 12 (0 self)
 Add to MetaCart
The Principle of Maximum Entropy is discussed and two classic probabilistic models of information retrieval, the Binary Independence Model of Robertson and Sparck Jones and the Combination Match Model of Croft and Harper are derived using the maximum entropy approach. The assumptions on which the classical models are based are not made. In their place, the probability distribution of maximum entropy consistent with a set of constraints is determined. It is argued that this subjectivist approach is more philosophically coherent than the frequentist conceptualization of probability that is often assumed as the basis of probabilistic modeling and that this philosophical stance has important practical consequences with respect to the realization of information retrieval research.
Monitoring UserSystem Performance in Interactive Retrieval Tasks
 PROC. RIAO 2004
, 2004
"... Monitoring usersystem performance in interactive search is a challenging task. Traditional measures of retrieval evaluation, based on recall and precision, are not of any use in real time, for they require a priori knowledge of relevant documents. This paper shows how a Shannon entropybased measur ..."
Abstract

Cited by 4 (1 self)
 Add to MetaCart
Monitoring usersystem performance in interactive search is a challenging task. Traditional measures of retrieval evaluation, based on recall and precision, are not of any use in real time, for they require a priori knowledge of relevant documents. This paper shows how a Shannon entropybased measure of usersystem performance naturally falls in the framework of (interactive) probabilistic information retrieval. The value of entropy of the distribution of probability of relevance associated with the documents in the collection can be used to monitor search progress in live testing, to allow for example the system to select an optimal combination of search strategies. User profiling and tuning parameters of retrieval systems are other important applications.
Probabilistic Information Retrieval Model for Dependency Structured Indexing System
 In Proceedings of the ACM SIGIRâ€™02 Workshop on Mathematical/Formal Methods in Information Retrieval, 2002. Proceedings of the Third NTCIR Workshop
, 2002
"... statistically independent from each another. However, independence assumption is obviously and openly understood to be wrong, so we present a new method of incorporating term dependence in probabilistic retrieval model by adapting a structural index system using dependency parse tree and the Chow Ex ..."
Abstract

Cited by 2 (1 self)
 Add to MetaCart
statistically independent from each another. However, independence assumption is obviously and openly understood to be wrong, so we present a new method of incorporating term dependence in probabilistic retrieval model by adapting a structural index system using dependency parse tree and the Chow Expansion to compensate the weakness of the assumption. In this paper, we describe a theoretic process to apply the Chow Expansion to the general probabilistic models and the stateoftheart 2Poisson model, and we reexamine the weight of phrase terms. Through the experiments on document collections, ETRIKEMONG in Korean, we demonstrate that the incorporation of term dependences using the Chow Expansion contribute to the improvement of performance in Probabilistic IR systems. Keywords term dependence, phrasal indexing, Chow Expansion, probabilistic model, 2Poisson model 1.
A Model for the Stopping Behavior of Users of Online Systems
, 1987
"... We examine a model in which the user of an online system continually updates his/her estimated probability of success, and quits or continues according to the expected utility of each action. The prior distribution of the unknown probability is a beta distribution, with mean determined by the a prio ..."
Abstract

Cited by 2 (1 self)
 Add to MetaCart
We examine a model in which the user of an online system continually updates his/her estimated probability of success, and quits or continues according to the expected utility of each action. The prior distribution of the unknown probability is a beta distribution, with mean determined by the a priori expectation of success, and variance determined by the confidence with which the user has that prior expectation. The stopping criterion depends upon the accumulated number of positive and negative reinforcements, and is a straight line in a suitable coordinate system.
To Daria ACKNOWLEDGMENTS
, 2004
"... This thesis would not have been possible without the help of my relatives, mentors and colleagues. First and foremost, I am thankful to my parents for igniting in me the passion for scientific discovery, for teaching me how to think and how to wonder. I was lucky to have a father who could explain t ..."
Abstract
 Add to MetaCart
This thesis would not have been possible without the help of my relatives, mentors and colleagues. First and foremost, I am thankful to my parents for igniting in me the passion for scientific discovery, for teaching me how to think and how to wonder. I was lucky to have a father who could explain trigonometry to a ten year old, and a mother who taught me to strive for perfection. I was also fortunate to have grandparents who served as role models of determination and hard work. I am truly indebted to John Stuber; without his kindness my life would have taken a very different path. I want to give special thanks to my wife for her understanding, patience and unending support during these difficult years. Having two academic advisors is a challenging but fascinating experience. I was fortunate to work closely with Bruce Croft, whose knowledge of the field is truly astounding, and who showed me the importance of seeing the wood behind the trees. I am also tremendously thankful to James Allan, who guided me since the early days of my undergraduate studies. I am particularly grateful to James for his willingness to give unbiased advice, even when it may have run contrary to his interests. I would like to thank Louise Knouse for turning the dry art of abstract mathematics into an accessible and fun discipline. I am also grateful to Donald Geman and Alexey Koloydenko for showing me the power and the aweinspiring beauty of probability theory. I am thankful to Sherre Myers for proofreading this work.
Computing Conditional Probabilities In Large Domains By Maximizing Renyi's Quadratic Entropy
, 2003
"... In this dissertation we discuss methods for efficiently approximating conditional probabilities in large domains by maximizing the entropy of the distribution given a set of constraints. The constraints are constructed from conditional probabilities, typically of loworder, that can be accurately co ..."
Abstract
 Add to MetaCart
In this dissertation we discuss methods for efficiently approximating conditional probabilities in large domains by maximizing the entropy of the distribution given a set of constraints. The constraints are constructed from conditional probabilities, typically of loworder, that can be accurately computed from the training data. By appropriately choosing the constraints, maximum entropy methods can balance the tradeoffs in errors due to bias and variance. Standard maximum entropy techniques are too computationally inefficient for use in large domains in which the set of variables that are being conditioned upon varies. Instead of using the standard measure of entropy first proposed by Shannon, we use a measure that lies within the family of R\'enyi's entropies. If we allow our probability estimates to occasionally lie outside the range from 0 to 1, we can efficiently maximize R\'enyi's quadratic entropy relative to the constraints using a set of linear equations. We develop two algorithms, the inverse probability method and recurrent linear network, for maximizing R\'enyi's quadratic entropy without bounds. The algorithms produce identical results. However, depending on the type of problem, one method may be more computationally efficient than the other. We also propose an extension to the algorithms for partially enforcing the constraints based on our confidence in them. Our algorithms are tested on several applications including: collaborative filtering, image retrieval and language modeling.
Distance, Minimum CrossEntropy, and Path methods. Background and Purpose of the Study
, 1988
"... The maximum entropy principle may be applied to the design of probabilistic retrieval systems. When there are inconsistent expert judgments, the resulting optimization problem cannot be solved. The inconsistency of the expert judgments can be revealed by solving a linear programming formulation. In ..."
Abstract
 Add to MetaCart
The maximum entropy principle may be applied to the design of probabilistic retrieval systems. When there are inconsistent expert judgments, the resulting optimization problem cannot be solved. The inconsistency of the expert judgments can be revealed by solving a linear programming formulation. In the case of inconsistent judgment, four plausible schemes are proposed in order to find revised judgments which are consistent with the true data structure but still reflect the original expert judgment. These schemes are the Interactive, Minimum