On feature relevance feedback methods : incorporating labeled user features

Graduate Thesis Or Dissertation

Public Deposited

Citeable URL: https://ir.library.oregonstate.edu/concern/graduate_thesis_or_dissertations/pc289m575

Descriptions

Attribute Name	Values
Creator	Oberst, Ian
Abstract	In text classification, labeling features is often less time consuming than labeling entire documents. In situations where very little labeled training data is available, feature relevance feedback has the potential to dramatically increase classification performance. We review previous work on incorporating feature relevance feedback in the form of labeled features and introduce a new method, the Feature Contrast Method, for using feature relevance feedback with locally weighted logistic regression. We show that our method is responsive to user feedback and significantly outperforms previously developed feature relevance feedback methods while remaining robust to noise in feature and training data. We also highlight several key issues that affect the performance of feature feedback methods.
License	All rights reserved
Resource Type	Masters Thesis
Date Available	2010-08-06T15:42:15+00:00
Date Issued	2010-06-07
Degree Level	Master's
Degree Name	Master of Science (M.S.)
Degree Field	Computer Science
Degree Grantor	Oregon State University
Commencement Year	2011
Advisor	Wong, Weng-Keen
Committee Member	Dietterich, Thomas Burnett, Margaret Lachenbruch, Peter
Academic Affiliation	Electrical Engineering and Computer Science
Non-Academic Affiliation	Oregon State University. Graduate School
Subject	Information retrieval
Rights Statement	In Copyright
Publisher	Oregon State University
Peer Reviewed	No
Language	English [eng]
Replaces	http://hdl.handle.net/1957/17411

Parents:

This work has no parents.

In Collection:

Thumbnail	Title	Date Uploaded	Visibility	Actions
	OberstIanS2011.pdf	2017-08-07	Public	Download