Please use this identifier to cite or link to this item:
Title: Natural two view learning
Authors: Dubey, Rachit.
Keywords: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence
Issue Date: 2012
Abstract: Co-training is a semi supervised learning method that effectively learns from a pool of labeled and unlabeled data and takes advantage of redundancy in the feature set. Co-training is known to work well when the assumptions it makes on the feature set hold true. In this project, we investigated the use of co-training for action recognition for the elderly people. Experimental results showed that co-training was able to boost the performance of action recognition with a very few number of labeled samples. In this project, we also present a new co-training strategy – natural two view learning which doesn’t require the prior existence of two redundant views. Our proposed strategy doesn’t require the feature set to be described with sufficient and redundant view and hence can be applied to a broader class of problems. The Experiments on UCI data sets indicates that the proposed natural two view learning algorithm improves classification accuracy especially when the number of labeled examples is very few.
Schools: School of Computer Engineering 
Research Centres: Centre for Computational Intelligence 
Rights: Nanyang Technological University
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:SCSE Student Reports (FYP/IA/PA/PI)

Files in This Item:
File Description SizeFormat 
  Restricted Access
2.76 MBAdobe PDFView/Open

Page view(s) 20

Updated on Jul 20, 2024


Updated on Jul 20, 2024

Google ScholarTM


Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.