Bootstrap prediction and Bayesian prediction under misspecified models

Fushiki, Tadayoshi

doi:10.3150/bj/1126126768

Cited by 21 publications

(24 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The bootstrap predictive distribution is obtained by applying the bagging algorithm to the plug-in distribution with the maximum likelihood estimator (MLE). The effectiveness of bootstrap prediction has been studied under the Kullback-Leibler loss Fushiki, 2005;Harris, 1989). It is known that Bayesian prediction is admissible when a proper prior is used (Aitchison, 1975).…”

Section: Introductionmentioning

confidence: 99%

Bayesian bootstrap prediction

Fushiki¹

2010

Journal of Statistical Planning and Inference

Self Cite

View full text Add to dashboard Cite

Section: Introductionmentioning

confidence: 99%

Bayesian bootstrap prediction

Fushiki¹

2010

Journal of Statistical Planning and Inference

Self Cite

View full text Add to dashboard Cite

“…However, the prediction by a learning machine θ usually has a bias and a variance, so that the predictive distribution given by (6) with no bias and the derived (7) seem to be inappropriate. So, we derive (5) from another approach.…”

Section: A Bayesian Bagging Predictionmentioning

confidence: 98%

“…One uses the hyper parameter applicable to several neural networks, such as perceptron and backpropagation networks, where the norm of the connection weights of the network is supposed to reflect the generalization performance and is used as the hyper parameter (see [2], [3]). Another uses the Bayesian bootstrap [4] enabling continuous weights of the data in the bootstrap sample sets instead of discrete weights for ordinary bootstrap, and its application to the bagging is called Bayesian bagging [5] or Bayesian bootstrap prediction [6],…”

Section: Introductionmentioning

confidence: 99%

Improving generalization performance of bagging ensemble via Bayesian approach

Kurogi

Harashima

2009

2009 IEEE International Symposium on Computational Intelligence in Robotics and Automation - (CIRA)

View full text Add to dashboard Cite

This paper describes a method for improving the generalization performance of bagging ensemble by means of using Bayesian approach. We examine the Bayesian prediction using bagging leaning machines for regression problems, and show a method to reduce the generalization loss defined by the square error of the prediction for test data. We examine and validate the effectiveness via numerical experiments using the CAN2s as learning machines, where the CAN2 is a neural net for learning efficient piecewise linear approximation of nonlinear functions.

show abstract

“…, y Z }, and use the expected model change to approximate the true model change. The use of bootstrap to estimate prediction distribution has been well investigated [35]. Thus, the final EMCM sampling function for linear regression is expressed as…”

Section: Algorithm 1 Emcm For Linear Regressionmentioning

confidence: 99%

Batch Mode Active Learning for Regression With Expected Model Change

Cai

Zhang

2017

IEEE Trans. Neural Netw. Learning Syst.

View full text Add to dashboard Cite

While active learning (AL) has been widely studied for classification problems, limited efforts have been done on AL for regression. In this paper, we introduce a new AL framework for regression, expected model change maximization (EMCM), which aims at choosing the unlabeled data instances that result in the maximum change of the current model once labeled. The model change is quantified as the difference between the current model parameters and the updated parameters after the inclusion of the newly selected examples. In light of the stochastic gradient descent learning rule, we approximate the change as the gradient of the loss function with respect to each single candidate instance. Under the EMCM framework, we propose novel AL algorithms for the linear and nonlinear regression models. In addition, by simulating the behavior of the sequential AL policy when applied for k iterations, we further extend the algorithms to batch mode AL to simultaneously choose a set of k most informative instances at each query time. Extensive experimental results on both UCI and StatLib benchmark data sets have demonstrated that the proposed algorithms are highly effective and efficient.

show abstract

Bootstrap prediction and Bayesian prediction under misspecified models

Cited by 21 publications

References 9 publications

Bayesian bootstrap prediction

Bayesian bootstrap prediction

Improving generalization performance of bagging ensemble via Bayesian approach

Batch Mode Active Learning for Regression With Expected Model Change

Contact Info

Product

Resources

About