The technology for building knowledge-based systems by inductive inference from examples has been demonstrated successfully in several practical applications. This paper summarizes an approach to synthesizing decision trees that has been used in a variety of systems, and it describes one such system, ID3, in detail. Results from recent studies show ways in which the methodology can be modified to deal with information that is noisy and/or incomplete. A reported shortcoming of the basic algorithm is discussed and two means of overcoming it are compared. The paper concludes with illustrations of current research directions.

/pdf/induction-of-decision-trees-f74o3tgil7.pdf

Induction of Decision Trees

The effect of outliers on reaction time analyses is evaluated. The first section assesses the power of different methods of minimizing the effect of outliers on analysis of variance (ANOVA) and makes recommendations about the use of transformations and cutoffs. The second section examines the effect of outliers and cutoffs on different measures of location, spread, and shape and concludes using quantitative examples that robust measures are much less affected by outliers and cutoffs than measures based on moments. The third section examines fitting explicit distribution functions as a way of recovering means and standard deviations and concludes that unless fitting the distribution function is used as a model of distribution shape, the method is probably not worth routine use.

Methods for dealing with reaction time outliers.

/pdf/induction-of-decision-trees-3ruvtu0emb.pdf

In this paper, we explore the relationship between real exchange rates and real interest rate differentials in the United States, Germany, Japan, and the United Kingdom. Contrary to theories based on the joint hypothesis that domestic prices are sticky and monetary disturbances are predominant, we find little evidence of a stable relationship between real interest rates and real exchange rates. We consider both in-sample and out-of-sample tests. One hypothesis that is consistent with our findings is that real disturbances (such as productivity shocks) may be a major source of exchange rate volatility. THIS PAPER INVESTIGATES THE empirical relationship between major currency real exchange rates and real interest rates over the modern (post-March 1973) flexible rate experience. The exchange rates examined here include the dollar/ mark, dollar/yen, and dollar/pound rates. Our two major findings are as follows. First, the data do not indicate a strong correspondence between real interest rate differentials (short-term or long-term) and real exchange rates. This finding appears to conflict with the predictions of most monetary and portfolio balance models of exchange rate determination, though the conflict can be substantially reconciled if aggregate disturbances are primarily real in nature (i.e., changes in productivity, tastes, etc.). It is true that in many cases the sign of the estimated exchange rate-interest rate differential relationship is consistent with the possible predominance of financial market disturbances, but the relationship is not stable enough to be statistically significant. Second, although one does find some evidence of a unit root in both real exchange rates and long-term (but not shortterm) real interest differentials, these two series do not appear to be linearly cointegrated. Hence, the nonstationarity (or near nonstationarity) in the two series cannot be attributed to the same factor. In Section I, we briefly describe a class of small-scale monetary models of exchange rate determination. The importance of this class of models for empirical work derives from its strong predictions about how the exchange rate will move

Was It Real? The Exchange Rate‐Interest Differential Relation over the Modern Floating‐Rate Period

Strategic management theories invoke the concept of competitive advantage to explain firm performance, and empirical research investigates competitive advantage and describes how it operates. But as a performance hypothesis, competitive advantage has received surprisingly little formal justification, particularly in light of its centrality in strategy research and practice. As it happens, the core hypothesis - that competitive advantage produces sustained superior performance - finds little support in formal deductive or inductive inference, and the leading theories of competitive advantage incorporate refutation barriers that preclude meaningful empirical tests. The logical and philosophical foundations of the competitive advantage hypothesis are explored, locating its philosophical foundations in the epistemologies of Bayesian induction, abductive inference and an instrumentalist, pragmatic philosophy of science.

/pdf/competitive-advantage-logical-and-philosophical-3n76b7cxed.pdf

Competitive advantage: logical and philosophical considerations

1. Probability and Distributions. 2. Multivariate Distributions. 3. Some Special Distributions. 4. Some Elementary Statistical Inferences 5. Consistency and Limiting Distributions 6. Maximum Likelihood Methods. 7. Sufficiency. 8. Optimal Tests of Hypotheses. 9. Inferences about Normal Models. 10. Nonparametric and Robust Statistics. 11. Bayesian Statistics. Appendix A. Mathematical Comments Appendix B. R-Functions. Appendix C. Tables of Distributions Appendix D. List of Common Distributions Appendix E. References Appendix F. Answers to Selected Exercises Index

Shirley Hitchcock

Papers

Introduction to Mathematical Statistics.