文档介绍：李群深层结构学****算法研究
中文摘要
李群机器学****与深层结构学****是近年来倍受人们关注的新的机器学****方法，本文
将这两种方法进行有机融合，给出了李群深层结构学****算法。主要包括以下几方面的
内容：
1） 分析了李群机器学****以及深层结构学****的研究现状；
2） 给出了构造深层结构的过程；
3） 给出了在深层结构中嵌入李群半监督学****的算法、深层结构的逐层学****算
法和启发式学****算法，并将其应用在 ORL 标准人脸数据库的识别中，给出
识别的精确度；
4） 通过实验验证，李群深层结构学****算法可以有效的识别人脸。
通过本文的研究，一方面丰富了李群机器学****的内容；另一方面将李群深层结构
学****算法应用在人脸识别系统上取得了明显效果。本文的研究证明了深层结构可以有
效地将复杂问题通过分层过程变成简单问题，为解决复杂问题提供了一种新的解决方
案。
关键词：李群机器学****李群半监督学****李群深层结构学****逐层学****启发式学<br****br/>
作 者：何文慧
指导老师：李凡长
I
Research on Lie Group Deep Structure Learning
Algorithm
Abstract
Lie group machine learning and deep structure learning have been taken much
attention in recent years as new machine learning methods. In this paper, we integrate these
two algorithms and put forward Lie group deep structure learning algorithm. The main
research results are as follows:
1） Introduction of Lie group machine learning and deep structure learning;
2） How to process deep structure;
3） A semi-supervised learning based on Lie group embedded in deep structure, the
layered learning algorithm and heuristic learning algorithm are proposed, and
their applications in the standard ORL database has been proved by the accuracy
of face recognition.
4） The experiments have show that Lie group deep structure learning algorithm can
effectively identify face.
On one hand, our research has enriched the content of Lie group machine learning. On
the other hand, it can be seen from our research that our proposed algorithms perform well
in human face identification and that deep structure can transform the complex problems
into simple ones, which provides a novel way to the complex problems.
Key words: Lie group, Lie gr