期刊文献+
共找到2篇文章
< 1 >
每页显示 20 50 100
Davis's Inequalities for Two-Parameter Banach Space Valued Strong Martingales and Geometrical Properties of Banach Spaces 被引量:3
1
作者 YE Chen GAN Shi xin ZHANG Feng 《Wuhan University Journal of Natural Sciences》 CAS 2000年第2期131-136,共6页
In this paper we investigated theL 1 norm inequalities of theP square and the maximal functions of two-parameterB-valued strong martingales, which can be applied to characterizep-smoothness andq-convexity of Banach sp... In this paper we investigated theL 1 norm inequalities of theP square and the maximal functions of two-parameterB-valued strong martingales, which can be applied to characterizep-smoothness andq-convexity of Banach spaces. 展开更多
关键词 two parameterB valued strong martingale Davis’s inequality p smoothable q convexifiable
下载PDF
Deep advantage learning for optimal dynamic treatment regime
2
作者 Shuhan Liang Wenbin Lu Rui Song 《Statistical Theory and Related Fields》 2018年第1期80-88,共9页
Recently deep learning has successfully achieved state-of-the-art performance on many difficulttasks. Deep neural networks allow for model flexibility and process features without the needof domain knowledge. Advantag... Recently deep learning has successfully achieved state-of-the-art performance on many difficulttasks. Deep neural networks allow for model flexibility and process features without the needof domain knowledge. Advantage learning (A-learning) is a popular method in dynamic treatment regime (DTR). It models the advantage function, which is of direct relevance to optimaltreatment decision. No assumptions on baseline function are made. However, there is a paucityof literature on deep A-learning. In this paper, we present a deep A-learning approach to estimate optimal DTR. We use an inverse probability weighting method to estimate the differencebetween potential outcomes. Parameter sharing of convolutional neural networks (CNN) greatlyreduces the amount of parameters in neural networks, which allows for high scalability. Convexified convolutional neural networks (CCNN) relax the constraints of CNN for optimisation purpose.Different architectures of CNN and CCNN are implemented for contrast function estimation.Both simulation results and application to the STAR*D (Sequenced Treatment Alternatives toRelieve Depression) trial indicate that the proposed methods outperform penalised least squareestimator. 展开更多
关键词 Advantage learning convexified convolutional neural networks convolutional neural networks dynamic treatment regime inverse probability weighting
原文传递
上一页 1 下一页 到第
使用帮助 返回顶部