Dissertation > Excellent graduate degree dissertation topics show
Research on Dimensionality Reduction Algorithms Based Locally Linear Analysis
Author: LiuShengLan
Tutor: ZuoDeQin
School: Liaoning Normal University
Course: Computer Software and Theory
Keywords: High-dimensional data dimensionality reduction Manifold Learning Local linearization Local tangent space
CLC: TP301.6
Type: Master's thesis
Year: 2011
Downloads: 54
Quote: 1
Read: Download Dissertation
Abstract
|
With the rapid development of the information , multimedia and digital technology , along with the arrival of the era of high-dimensional data and has become a powerful tool in the description of the objective world , such as gene expression , video tracking , medical image processing , high -dimensional time series analysis , At the same time , the traditional classification, clustering algorithm can not be applied in the processing of high-dimensional data , there is an urgent need to find a data dimensionality reduction method , the emergence of high-dimensional data manifold learning dimensionality reduction provides a very good way . Manifold learning in more than 10 years of development , the efforts of many scholars at home and abroad , has begun to mature , and emerged much to learn . For example : isometric mapping , local linear embedded (LLE) into the local tangent space alignment (LTSA) algorithm . LLE and LTSA are based on the local approximate linearization assumptions and proposed nonlinear dimensionality reduction method can get better results in the real world of high-dimensional data . But in many cases , the local data there is often a high curvature distribution and noise , the local method is very sensitive to the above situation , this time LLE and LTSA will not be able to get the correct low- dimensional embedding manifold learning how to solve such problems become an important branch . This article is mainly for the important issues in the above manifold learning solution : ( 1 ) analysis of the geometric properties of the local tangent space , on this basis propose an adaptive neighborhood selection methods , and the LTSA algorithms be improved . ( 2 ) analysis of the impact of noise and high curvature of the low-dimensional space , and noise classification proposed an anti- noise ability angles global embedding algorithm . (3) LLE algorithm , for example , local linearization problem discussed , given an approximate linear standard , while in the source data is the case of sparsely distributed , give an analysis based on sparse embedded dimensionality reduction method . Finally, the experiments confirmed the effectiveness of the proposed method .
|
Related Dissertations
- The Classification of High Dimsnsion Flew Field Based on Manifold Learning,V231.3
- Feature Extraction, Selection and Combination in Lipreading,TP391.41
- Face Recognition Based on Statistical Learning Methods,TP391.41
- Reasearch on the Fiber Bundle Learning Algorithms Based on Manifold Learning,TP301.6
- Manifold Learning Methods for Hyperspectral Image Classification and Anomaly Detection,TP391.41
- Research on Semantic-based Web Image Classification,TP391.41
- Research on Tensor Subspace Face Recognition Algorithm,TP391.41
- Study of Data Reduction Technique Based on Manifold Learning,TP311.13
- Research on Method of Rotating Machinery Fault Diagnosis Based on Manifold Learning,TH165.3;O186.12
- Several manifold learning about the basic problems with the core algorithm,TP311.13
- The Application of Manifold-Learning Algorithms in Pattern Recognition,TP391.4
- Based on manifold learning of supervised dimensionality reduction method,TP391.41
- Research on Finger Vein Recognition Based on Manifold Learning and Extension Classifier,TP391.41
- Research on Methods of Face Recognition Based on Manifold Studying Subspace,TP391.41
- Video Copy Detection Based on Robust Hashing,TP391.41
- Modeling and Control of MLCR-CSC Based on STATCOM,TM76
- Based on features preserved rare hyperspectral image analysis dimension,TP751
- Novel Intelligent Pattern Recognition Methods and Its Application,TP391.4
- The Research on Face Feature Extraction Based on Manifold Learning,TP391.41
- Research on Dimensionality Reduction Algorithms and Its Applications,TP391.41
CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > General issues > Theories, methods > Algorithm Theory
© 2012 www.DissertationTopic.Net Mobile
|