scispace - formally typeset
Search or ask a question

Showing papers by "William G. Macready published in 1999"


Journal ArticleDOI
TL;DR: This paper presents several techniques for estimating the generalization error of a bagged learning algorithm without invoking yet more training of the underlying learning algorithm (beyond that of the bagging itself), as is required by cross-validation-based estimation.
Abstract: Bagging (Breiman, 1994a) is a technique that tries to improve a learning algorithm‘s performance by using bootstrap replicates of the training set (Efron & Tibshirani, 1993, Efron, 1979). The computational requirements for estimating the resultant generalization error on a test set by means of cross-validation are often prohibitive, for leave-one-out cross-validation one needs to train the underlying algorithm on the order of mν times, where m is the size of the training set and ν is the number of replicates. This paper presents several techniques for estimating the generalization error of a bagged learning algorithm without invoking yet more training of the underlying learning algorithm (beyond that of the bagging itself), as is required by cross-validation-based estimation. These techniques all exploit the bias-variance decomposition (Geman, Bienenstock & Doursat, 1992, Wolpert, 1996). The best of our estimators also exploits stacking (Wolpert, 1992). In a set of experiments reported here, it was found to be more accurate than both the alternative cross-validation-based estimator of the bagged algorithm‘s error and the cross-validation-based estimator of the underlying algorithm‘s error. This improvement was particularly pronounced for small test sets. This suggests a novel justification for using bagging—more accurate estimation of the generalization error than is possible without bagging.

145 citations


Patent
01 Jul 1999
TL;DR: In this article, the authors present a comprehensive system and method for operations management which has the reliability and adaptability to handle failures and changes respectively within the economic environment, using technology graphs, landscape representations and automated markets.
Abstract: The present invention presents a comprehensive system and method for operations management which has the reliability and adaptability to handle failures and changes respectively within the economic environment. The present invention presents a framework of features which include technology graphs, landscape representations and automated markets to achieve the requisite reliability and adaptability.

99 citations


Patent
22 Dec 1999
TL;DR: This article presented a system and method of economic analysis and prediction which dynamically adapts to a changing economic environment by selecting or synthesizing an economic model from a set of economic models based on the selected model's ability to make accurate predictions about an actual economic market.
Abstract: The present invention presents a system and method of economic analysis and prediction which dynamically adapts to a changing economic environment by selecting or synthesizing an economic model from a set of economic models based on the selected model's ability to make accurate predictions about an actual economic market. Specifically, the method and system of the present invention forms a space of different economic models (105), forms a behavorial landscape by extracting observables from executions of the economic models (120), and performs model selection and composite model synthesis through optimization (130) over the behavioral landscape.

40 citations


Posted Content
TL;DR: In this paper, the authors introduce a technology landscape into the search-theoretic framework and show that early in the search for technological improvements, if the initial position is poor or average, it is optimal to search far away on the technology landscape; but as the firm succeeds in finding technological improvements it was optimal to confine search to a local region of the landscape.
Abstract: Technological change at the firm-level has commonly been modeled as random sampling from a fixed distribution of possibilities. Such models typically ignore empirically important aspects of the firm's search process, namely the related observations that a firm's current technology constrains future innovation and that firms' technological search tends to be local in nature. In this paper we explicitly treat these aspects of the firm's search for technological improvements by introducing a technology landscape into the search-theoretic framework. Technological search is modeled as movement over a technology landscape with the firm's adaptive walk constrained by the firm's location on the landscape, the correlation structure of the landscape and the cost of innovation. We show that the standard search model is attained as a limiting case of a more general landscape search model. We obtain two key results, otherwise unavailable in the standard search model: the presence of local optima in space of technological possibilities and the determination of the optimal search distance. We find that early in the search for technological improvements, if the initial position is poor or average, it is optimal to search far away on the technology landscape; but as the firm succeeds in finding technological improvements it is optimal to confine search to a local region of the landscape. Notably, we obtain diminishing returns to search without having to make the assumption that the firm's repeated draws from the search space are independent and identically distributed. The distinction between dramatic technological improvements ("innovations") and minor technological improvements hinges on the distance at which a firm decides to sample the technology landscape. Submitted to J. Pol. Econ.

19 citations


Patent
31 Aug 1999
TL;DR: In this paper, the authors model the search process of a firm's search for technological improvements as a Markov random field, and show that early in the search, if the initial position is poor or average, it is optimal to search far away on the technology landscape; but as the firm succeeds in finding technological improvements, and confine search to a local region of the landscape, the optimal strategy is to assign a reservation price to each possible technology.
Abstract: Technological change at the firm-level has commonly been modeled as random sampling from a fixed distribution of possibilities. Such models, however, typically ignore empirically important aspects of the firm's search process, notably the observation that the present state of the firm guides future innovation. In this paper we explicitly treat this aspect of the firm's search for technological improvements by introducing a “technology landscape” into an otherwise standard dynamic programming setting where the optimal strategy is to assign a reservation price to each possible technology. Search is modeled as movement, constrained by the cost of innovation, over the technology landscape. Simulations are presented on a stylized technology landscape while analytic results are derived using landscapes that are similar to Markov random fields. We find that early in the search for technological improvements, if the initial position is poor or average, it is optimal to search far away on the technology landscape; but as the firm succeeds in finding technological improvements it is optimal to confine search to a local region of the landscape. We obtain the result that there are diminishing returns to search without having to make the assumption that the firm's repeated draws from the search space are independent and identically distributed.

4 citations