Powered partial least squares discriminant analysis
Powered partial least squares discriminant analysis
复制标题
DOI：
10.1002/cem.1186
复制
发表时间：
2009-01-01
影响因子：
2.4
通讯作者：
Indahl, Ulf Geir
中科院分区：
文献类型：
--
作者：
Liland, Kristian Hovde;Indahl, Ulf Geir
从PLS- da的基本部分，Fisher的典型判别分析（FCDA）和动力PLS （PPLS）出发，我们提出了动力PLS的概念（PPLS- da）。通过利用一系列减少线性变换的数据（与普通PLS-DA组件的计算一致），PPLS-DA通过最大化与FCDA相关的参数化瑞利商，从转换后的数据中计算每个组件。由有动力的PILS方法发现的模型有助于揭示特定预测因子的相关性，并且通常需要比普通的PILS对应物更少和更简单的组件。从对可用于优化的功率施加限制的可能性出发，我们获得了一种传统PILS方法无法获得的预测建模的探索性方法。版权所有(C) 2008约翰威利父子有限公司
From the fundamental parts of PLS-DA, Fisher's canonical discriminant analysis (FCDA) and Powered PLS (PPLS), we develop the concept of powered PILS for classification problems (PPLS-DA). By taking advantage of a sequence of data reducing linear transformations (consistent with the computation of ordinary PLS-DA components), PPLS-DA computes each component from the transformed data by maximization of a parameterized Rayleigh quotient associated with FCDA. Models found by the powered PILS methodology can contribute to reveal the relevance of particular predictors and often requires fewer and simpler components than their ordinary PILS counterparts. From the possibility of imposing restrictions on the powers available for optimization we obtain an explorative approach to predictive modeling not available to the traditional PILS methods. Copyright (C) 2008 John Wiley & Sons, Ltd.