KReSIT Logo Kanwal Rekhi School of Information Technology
IITB Logo
IIT Bombay
 
 Research
 Groups
 Publications
 Projects
 Student Projects
 Ongoing Projects
 Past Projects
 Proposed Projects-2005
 Projects Archives
 ASL BE/MCA Projects
 Sponsored Projects
 Seminars
 Labs
 Techtalks
 

Home > Research > Projects > Students > Previous 

Students' Projects


Statistical Learners for Information Extraction:An Empiricial Approach

Kaushal Mittal

Abstract


Information Extraction from unstructured data available on the web and other sources is one of the interesting problems in machine learning. Generative models like Hidden Markov Model (HMM) and conditional models like Maximum Entropy Markov Models (MEMM),Conditional Random Fields (CRF) and their variants, have been used for information extraction.Lot of variations have been proposed on the basis of training algorithms, regularization techniques, loss functions etc. Most of the work has compared the models in terms of their accuracies. Our work focus on comparing the models by comparing their parameter vectors and gain insight over the benefits of one model over the another.We propose to develop approaches for analyzing and comparing the models, grouping the features, performing sensitivity analysis of the conditional models and hence finding the minimum set of features whose weights can be changed to improve the accuracy significantly.We propose to design and develop a workbench to support these approaches and facilitate better understanding about the features and the models. In this report we discuss our approach for feature grouping and model comparisons.








Printer friendly    Comment
  Copyright © 2004 KReSIT, IIT Bombay. All rights reserved sitemap    
  Kanwal Rekhi School of Information Technology, Indian Institute of Technology Bombay, Powai, Mumbai - 400 076.
+91-22-2576 7901/02. Fax: +91-22-2572 0022
Designed by Kamlesh