Results 1  10
of
263
Inference with an incomplete model of english auctions
 JOURNAL OF POLITICAL ECONOMY
, 2003
"... While English auctions are the most common in practice, their rules typically lack sufficient structure to yield a tractable theoretical model without significant abstractions. Rather than relying on one stylized model to provide an exact interpretation of the data, we explore an incomplete model ba ..."
Abstract

Cited by 155 (7 self)
 Add to MetaCart
While English auctions are the most common in practice, their rules typically lack sufficient structure to yield a tractable theoretical model without significant abstractions. Rather than relying on one stylized model to provide an exact interpretation of the data, we explore an incomplete model based on two simple assumptions: bidders neither bid more than their valuations nor let an opponent win at a price they would be willing to beat. Focusing on the symmetric independent private values paradigm, we show that this limited structure enables construction of informative bounds on the distribution function characterizing bidder demand, on the optimal reserve price, and on the effects of observable covariates on bidder valuations. If the standard theoretical model happens to be the true model, our bounds collapse to the true features of interest. In contrast, when the true datagenerating process deviates in seemingly small ways from that implied by equilibrium in the standard theoretical model, existing methods can yield misleading results that need not even lie within our bounds. We report results from Monte Carlo experiments illustrating the performance of our approach and comparing it to others. We apply our ap
A Comparison of Selection Schemes used in Genetic Algorithms
 Gloriastrasse 35, CH8092 Zurich: Swiss Federal Institute of Technology (ETH) Zurich, Computer Engineering and Communications Networks Lab (TIK
, 1995
"... Genetic Algorithms are a common probabilistic optimization method based on the model of natural evolution. One important operator in these algorithms is the selection scheme for which a new description model is introduced in this paper. With this a mathematical analysis of tournament selection, trun ..."
Abstract

Cited by 95 (3 self)
 Add to MetaCart
Genetic Algorithms are a common probabilistic optimization method based on the model of natural evolution. One important operator in these algorithms is the selection scheme for which a new description model is introduced in this paper. With this a mathematical analysis of tournament selection, truncation selection, linear and exponential ranking selection and proportional selection is carried out that allows an exact prediction of the fitness values after selection. The further analysis derives the selection intensity, selection variance, and the loss of diversity for all selection schemes. For completion a pseudocode formulation of each method is included. The selection schemes are compared and evaluated according to their properties leading to an unified view of these different selection schemes. Furthermore the correspondence of binary tournament selection and ranking selection in the expected fitness distribution is proven. Foreword This paper is the revised and extended versio...
A Comparison of Selection Schemes used in Evolutionary Algorithms
 Evolutionary Computation
, 1997
"... Evolutionary Algorithms are a common probabilistic optimization method based on the model of natural evolution. One important operator in these algorithms is the selection scheme for which in this paper a new description model based on fitness distributions is introduced. ..."
Abstract

Cited by 81 (2 self)
 Add to MetaCart
Evolutionary Algorithms are a common probabilistic optimization method based on the model of natural evolution. One important operator in these algorithms is the selection scheme for which in this paper a new description model based on fitness distributions is introduced.
Linear and Order Statistics Combiners for Pattern Classification
 Combining Artificial Neural Nets
, 1999
"... Several researchers have experimentally shown that substantial improvements can be obtained in difficult pattern recognition problems by combining or integrating the outputs of multiple classifiers. This chapter provides an analytical framework to quantify the improvements in classification resul ..."
Abstract

Cited by 74 (8 self)
 Add to MetaCart
(Show Context)
Several researchers have experimentally shown that substantial improvements can be obtained in difficult pattern recognition problems by combining or integrating the outputs of multiple classifiers. This chapter provides an analytical framework to quantify the improvements in classification results due to combining. The results apply to both linear combiners and order statistics combiners. We first show that to a first order approximation, the error rate obtained over and above the Bayes error rate, is directly proportional to the variance of the actual decision boundaries around the Bayes optimum boundary. Combining classifiers in output space reduces this variance, and hence reduces the "added" error. If N unbiased classifiers are combined by simple averaging, the added error rate can be reduced by a factor of N if the individual errors in approximating the decision boundaries are uncorrelated. Expressions are then derived for linear combiners which are biased or correlated, and the effect of output correlations on ensemble performance is quantified. For order statistics based nonlinear combiners, we derive expressions that indicate how much the median, the maximum and in general the ith order statistic can improve classifier performance. The analysis presented here facilitates the understanding of the relationships among error rates, classifier boundary distributions, and combining in output space. Experimental results on several public domain data sets are provided to illustrate the benefits of combining and to support the analytical results.
Evolutionary Algorithms in Noisy Environments: Theoretical Issues and Guidelines for Practice
 Computer Methods in Applied Mechanics and Engineering
, 1998
"... This paper is devoted to the effects of fitness noise in EAs (evolutionary algorithms). After a short introduction to the history of this research field, the performance of GAs (genetic algorithms) and ESs (evolution strategies) on the hypersphere test function is evaluated. It will be shown that t ..."
Abstract

Cited by 68 (6 self)
 Add to MetaCart
(Show Context)
This paper is devoted to the effects of fitness noise in EAs (evolutionary algorithms). After a short introduction to the history of this research field, the performance of GAs (genetic algorithms) and ESs (evolution strategies) on the hypersphere test function is evaluated. It will be shown that the main effects of noise  the decrease of convergence velocity and the residual location error R1  are observed in both GAs and ESs.
ICA Using Spacings Estimates of Entropy
 Journal of Machine Learning Research
, 2003
"... This paper presents a new algorithm for the independent components analysis (ICA) problem based on an efficient entropy estimator. Like many previous methods, this algorithm directly minimizes the measure of departure from independence according to the estimated KullbackLeibler divergence betwee ..."
Abstract

Cited by 68 (3 self)
 Add to MetaCart
(Show Context)
This paper presents a new algorithm for the independent components analysis (ICA) problem based on an efficient entropy estimator. Like many previous methods, this algorithm directly minimizes the measure of departure from independence according to the estimated KullbackLeibler divergence between the joint distribution and the product of the marginal distributions. We pair this approach with efficient entropy estimators from the statistics literature. In particular, the entropy estimator we use is consistent and exhibits rapid convergence. The algorithm based on this estimator is simple, computationally efficient, intuitively appealing, and outperforms other well known algorithms. In addition, the estimator's relative insensitivity to outliers translates into superior performance by our ICA algorithm on outlier tests. We present favorable comparisons to the Kernel ICA, FASTICA, JADE, and extended Infomax algorithms in extensive simulations. We also provide public domain source code for our algorithms.
Identification of Standard Auction Models
, 2001
"... We present new identification results for models of firstprice, secondprice, ascending (English), and descending (Dutch) auctions. We analyze a general specification of the latent demand and information structure, nesting as special cases the pure private values and pure common values models, and ..."
Abstract

Cited by 63 (6 self)
 Add to MetaCart
We present new identification results for models of firstprice, secondprice, ascending (English), and descending (Dutch) auctions. We analyze a general specification of the latent demand and information structure, nesting as special cases the pure private values and pure common values models, and allowing both ex ante symmetric and asymmetric bidders. We address identification of a series of nested models and derive testable restrictions that enable discrimination between models on the basis of observed data. The simplest model–that of symmetric independent private values–is nonparametrically identified even if only the transaction price from each auction is observed. For more complex models, identification and testable restrictions are obtained when additional information of one or more of the following types is available: (i) the identity of the winning bidder or other bidders, (ii) one or more bids in addition to the transaction price; (iii) exogenous variation in the number of bidders; (iv) bidderspecific covariates; (v) auctionspecific covariates. While many private values (PV) models are nonparametrically
Nonparametric estimation of an eBay auction model with an unknown number of bidders
, 2004
"... In this paper, I present new identification results and proposes an estimation method for an eBay auction model with an application. A key difficulty with data from eBay auctions is the fact that the number of potential bidders willing to pay the reserve price is not observable and the number of pot ..."
Abstract

Cited by 47 (1 self)
 Add to MetaCart
(Show Context)
In this paper, I present new identification results and proposes an estimation method for an eBay auction model with an application. A key difficulty with data from eBay auctions is the fact that the number of potential bidders willing to pay the reserve price is not observable and the number of potential bidders varies auction by auction. While this precludes application of existing estimation methods, I show that this need not preclude structural analysis of the available bid data. In particular, I show that within the symmetric independent private values (IPV) model, observation of any two valuations of which rankings from the top is known (for example, the second and thirdhighest valuations) nonparametrically identifies the bidders ' underlying value distribution. In contrast to the results of previous studies, the researcher does not need to know the number of potential bidders willing to pay the reserve price nor assume that the number of potential bidders is fixed across auctions. I then propose a consistent estimator using the seminonparametric maximum likelihood estimation method developed by Gallant and his coauthors. Several Monte Carlo experiments are conducted to illustrate its performance. The simulation results show that the proposed estimator performs well. I apply the proposed method to university yearbook sales on eBay. Using my estimate of bidders ' value distribution, I explore the effects of sellers ' ratings on bidders ' value distribution; compute consumers' surplus; and examine a regularity assumption that is often made in the mechanism design literature.
On SelfAdaptive Features in RealParameter Evolutionary Algorithms
, 2001
"... Due to the flexibility in adapting to different fitness landscapes, selfadaptive evolutionary algorithms (SAEAs) have been gaining popularity in the recent past. In this paper, we postulate the properties that SAEA operators should have for successful applications in realvalued search spaces. Sp ..."
Abstract

Cited by 42 (7 self)
 Add to MetaCart
Due to the flexibility in adapting to different fitness landscapes, selfadaptive evolutionary algorithms (SAEAs) have been gaining popularity in the recent past. In this paper, we postulate the properties that SAEA operators should have for successful applications in realvalued search spaces. Specifically, population mean and variance of a number of SAEA operators, such as various realparameter crossover operators and selfadaptive evolution strategies, are calculated for this purpose. Simulation results are shown to verify the theoretical calculations. The postulations and population variance calculations explain why selfadaptive GAs and ESs have shown similar performance in the past and also suggest appropriate strategy parameter values which must be chosen while applying and comparing different SAEAs.