Regression methods aim at inducing models of numeric data. While most state-of-the-art machine learning methods for regression focus on inducing piecewise regression models (regression and model trees), we investigate the predictive performance of regression models based on polynomial equations. We present Ciper, an efficient method for inducing polynomial equations and empirically evaluate its...

As processor clock frequencies become faster, architecture-level design is becoming increasingly limited by factors such as on-chip variation. Parameter variation occurs in integrated circuits as the result of a variety of manufacturing and physical factors. In this paper, we examine the degree to which there is parameter variation on an FPGA. Data were gathered from a combinatorial logic devic...

Unit-root and variance-ratio tests are used to examine the trend properties and degree of persistence of industrial production in U.S. industries and comparable aggregates during the post World War I1 period. The evidence from unit-root tests suggests that less than one-half of these industries have output which may be characterized as a random walk. The variance-ratio test results generally su...

This paper studies inference on fixed effects in a linear regression model estimated from network data. We derive bounds on the variance of the fixed-effect estimator that uncover the importance of the smallest non-zero eigenvalue of the (normalized) Laplacian of the network and of the degree structure of the network. The eigenvalue is a measure of connectivity, with smaller values indicating l...