Discovery of Relevant Weights by Minimizing Cross-Validation Error
نویسندگان
چکیده
In order to discover relevant weights of neural networks, this paper proposes a novel method to learn a distinct squared penalty factor for each weight as a minimization problem over the cross-validation error. Experiments showed that the proposed method works well in discovering a polynomial-type law even from data containing irrelevant variables and a small amount of noise.
منابع مشابه
Cross-validation of matching correlation analysis by resampling matching weights
The strength of association between a pair of data vectors is represented by a nonnegative real number, called matching weight. For dimensionality reduction, we consider a linear transformation of data vectors, and define a matching error as the weighted sum of squared distances between transformed vectors with respect to the matching weights. Given data vectors and matching weights, the optima...
متن کاملMinimizing the Weights dispersion in Cross-Efficiency Measurement in data envelopment analysis
Because of the piecewise linear nature of the data envelopment analysis (DEA) frontier, the optimal multipliers of the DEA models may not be unique. Choosing weights from alternative optimal solutions of dual multiplier models is one of the most frequently studied subjects in the context of DEA. In this paper, the authors have been inspired by the idea of Cooper et al. (2011) to propose a li...
متن کاملJackknife Model Averaging for Quantile Regressions
In this paper we consider the problem of frequentist model averaging for quantile regression (QR) when all the models under investigation are potentially misspecified and the number of parameters in some or all models is diverging with the sample size To allow for the dependence between the error terms and the regressors in the QR models, we propose a jackknife model averaging (JMA) estima...
متن کاملLearning Chemical Trends in Heterogeneous Catalysis
Despite tremendous advances in the understanding of catalysis in the past few decades, the discovery of improved materials for energy transformation processes has been guided by physical intuition and trial-and-error approaches. Even though computational simulations have provided a more rational approach, even the most systematic studies involving the most well-described systems require a large...
متن کاملOptimal convex error estimators for classification
A cross-validation error estimator is obtained by repeatedly leaving out some data points, deriving classifiers on the remaining points, computing errors for these classifiers on the left-out points, and then averaging these errors. The 0.632 bootstrap estimator is obtained by averaging the errors of classifiers designed from points drawn with replacement and then taking a convex combination of...
متن کامل