Results 1 
9 of
9
Empirical Analysis of Predictive Algorithm for Collaborative Filtering
 Proceedings of the 14 th Conference on Uncertainty in Artificial Intelligence
, 1998
"... 1 ..."
Dependency networks for inference, collaborative filtering, and data visualization
 Journal of Machine Learning Research
"... We describe a graphical model for probabilistic relationshipsan alternative tothe Bayesian networkcalled a dependency network. The graph of a dependency network, unlike aBayesian network, is potentially cyclic. The probability component of a dependency network, like aBayesian network, is a set of ..."
Abstract

Cited by 156 (10 self)
 Add to MetaCart
We describe a graphical model for probabilistic relationshipsan alternative tothe Bayesian networkcalled a dependency network. The graph of a dependency network, unlike aBayesian network, is potentially cyclic. The probability component of a dependency network, like aBayesian network, is a set of conditional distributions, one for each nodegiven its parents. We identify several basic properties of this representation and describe a computationally e cient procedure for learning the graph and probability components from data. We describe the application of this representation to probabilistic inference, collaborative ltering (the task of predicting preferences), and the visualization of acausal predictive relationships.
Improving Big Plans
 In Proceedings of the Fifteenth National Conference on Artificial Intelligence (AAAI98
, 1998
"... Past research on assessing and improving plans in domains that contain uncertainty has focused on analytic techniques that are exponential in the length of the plan. Little work has been done on choosing from among the many ways in which a plan can be improved. We present the Improve algorithm which ..."
Abstract

Cited by 19 (6 self)
 Add to MetaCart
Past research on assessing and improving plans in domains that contain uncertainty has focused on analytic techniques that are exponential in the length of the plan. Little work has been done on choosing from among the many ways in which a plan can be improved. We present the Improve algorithm which simulates the execution of large, probabilistic plans. Improve runs a data mining algorithm on the execution traces to pinpoint defects in the plan that most often lead to plan failure. Finally, Improve applies qualitative reasoning and plan adaptation algorithms to modify the plan to correct these defects. We have tested Improve on plans containing over 250 steps in an evacuation domain, produced by a domainspecific scheduling routine. In these experiments, the modified plans have over a 15% higher probability of achieving their goal than the original plan. Introduction Large, complex domains call for large, robust plans. However, today's stateoftheart planning algorithms cannot eff...
Seer: Maximum Likelihood Regression for LearningSpeed Curves
 University of Illinois at
, 1995
"... The research presented here focuses on modeling machinelearning performance. The thesis introduces Seer, a system that generates empirical observations of classificationlearning performance and then uses those observations to create statistical models. The models can be used to predict the number ..."
Abstract

Cited by 10 (0 self)
 Add to MetaCart
The research presented here focuses on modeling machinelearning performance. The thesis introduces Seer, a system that generates empirical observations of classificationlearning performance and then uses those observations to create statistical models. The models can be used to predict the number of training examples needed to achieve a desired level and the maximum accuracy possible given an unlimited number of training examples. Seer advances the state of the art with 1) models that embody the best constraints for classification learning and most useful parameters, 2) algorithms that efficiently find maximumlikelihood models, and 3) a demonstration on realworld data from three domains of a practicable application of such modeling. The first part of the thesis gives an overview of the requirements for a good maximumlikelihood model of classificationlearning performance. Next, reasonable design choices for such models are explored. Selection among such models is a task of nonlinear programming, but by exploiting appropriate problem constraints, the task is reduced to a nonlinear regression task that can be solved with an efficient iterative algorithm. The latter part of the thesis describes almost 100 experiments in the domains of soybean disease, heart disease, and audiological problems. The tests show that Seer is excellent at characterizing learningperformance and that it seems to be as good as possible at predicting learning
Landscape Ecology vol. 8 no. 1 pp 2537 (1993)
"... We tested the effects of increased landscape corridor width and corridor presence on the population dynamics and home range use of the meadow vole (Microtuspennsylvanicus) within a smallscale fragmented land scape. Our objective was to observe how populations behaved in patchy landscapes where th ..."
Abstract
 Add to MetaCart
We tested the effects of increased landscape corridor width and corridor presence on the population dynamics and home range use of the meadow vole (Microtuspennsylvanicus) within a smallscale fragmented land scape. Our objective was to observe how populations behaved in patchy landscapes where the animals home range exceeded or equaled patch size. We used a smallscale replicated experiment consisting of three sets of two patches each, unconnected or interconnected by lm or 5m widecorridors, established in an oldfield community (S.W. Ohio). Control (Om) treatments supported significantly lower vole densities than either corridor treatment. Females were the dominant resident sex establishing smaller home ranges (< than males Significantly more male voles dispersed between patches with corridors than between patches without corridors. However, no difference was observed regarding the number of male voles dispers ing between patches connected by corridors when compared to the number dispersing across treatments. Dis persal between connected patches was restricted to corridors based on tracking tube data. Corridor presence was more important than corridor width regarding the movement of male voles within their home range. 1.
INTERPOLATION AND INTERPRETATION FOR FIELD SCALE DECISION MAKING
"... Abstract: Yield mapping combine harvesters are capable of the continuous mapping of cereal yields. Explanations for the observed variability are frequently based upon the interpolation of soil data from a limited number of point samples. The point soil data and yield surface are integrated using the ..."
Abstract
 Add to MetaCart
Abstract: Yield mapping combine harvesters are capable of the continuous mapping of cereal yields. Explanations for the observed variability are frequently based upon the interpolation of soil data from a limited number of point samples. The point soil data and yield surface are integrated using the algorithms provided within mapping software and geographical information systems (GIS) that are increasingly available to the nonspecialist in spatial analyses. The surfaces and correlations produced are particularly sensitive to the choice and parameterisation of the algorithms, much of which is done within a ‘blackbox ’ environment. Furthermore, it is rarely established that spatial autocorrelation exists at the scale of sampling. Hence, interpolation may not always be valid. This can have a significant impact on the interpretation of the data and the process of field scale decision making. This paper discusses the consequences of using different interpolation algorithms and methods of visualisation, using examples from agricultural field consultancy within the United Kingdom. The authors conclude that a number of treatment classes should be defined for a few zones, rather than attempting to continuously vary application rates in response to an inaccurate model of the soil nutrient surface.
Dominican University
"... Using VITA service learning experiences to teach hypothesis testing and pvalue analysis ..."
Abstract
 Add to MetaCart
Using VITA service learning experiences to teach hypothesis testing and pvalue analysis
Perception & Psvc.hophysrc..\
"... Observers saw 234 different pairs of stochastically organized dot patterns and indicated which of the two patterns appeared to be more numerous. All of the data can be accounted for by supposing that the choice of the more numerous pattern is based on the determination of the occupancy indices of bo ..."
Abstract
 Add to MetaCart
Observers saw 234 different pairs of stochastically organized dot patterns and indicated which of the two patterns appeared to be more numerous. All of the data can be accounted for by supposing that the choice of the more numerous pattern is based on the determination of the occupancy indices of both patterns. Each dot is posited to have an impact upon its neighborhood in a constant occupancy radius R. The area of the stimulus plane occupied collectively by all dots provides a basis for judging relative numerosity; the pattern with the larger occupancy value is chosen as more numerous. The occupancy model, besides providing a general explanation of known numerosity illusions in strictly quantitative terms, can explain some puzzling aspects of numerosity perception. Quantification is one of the most impressive acts of the human mind. On many occasions, however, the direct onebyone counting of items is impossible: the number of objects is too large, the viewing time is too limited, the separation of alreadycounted objects from notyetcounted ones is too difficult, and so forth. Nevertheless,
ESfIMATING FUNCfIONALS OF ONEDIMENSIONAL GIBBS STATES by
"... Some estimators of maximum likelihood type are constructed for estimating functionals of one'dimensional Gibbs states. We also show that those estimators are strongly consistent. asymptotically normal 'and asymptotically efficient. Key words and phrases: ..."
Abstract
 Add to MetaCart
Some estimators of maximum likelihood type are constructed for estimating functionals of one'dimensional Gibbs states. We also show that those estimators are strongly consistent. asymptotically normal 'and asymptotically efficient. Key words and phrases: