ODIEProject

=Semantic Annotation Framework=

==

Project Team:
[|Quratulain Rajput] [|Sajjad Haider]

Institute:
[|Artificial Intelligence Lab], [|Faculty of Computer Science, IBA]

**Objective:**
BNOSA ( **B** ayesian **N**etwork and **O** ntology based **S** emantic **A** nnotation) is a tool for annotating web documents with semantic web content. The purpose of this tool is to extract the required information from unstructured and ungrammatical data sources and associate semantic tags with them. It uses extraction ontology to conceptualize a problem domain and use Bayesian network for conflict resolution and missing value prediction. This framework implements a new algorithm for ontology based information extraction from web pages, further this approach is scalable to handle variety of domain information by dynamically plug the domain specific ontology to the application.

Development Tools:
[|Protege] [|Jena ontology API] [|JAVA] [|Belief Network Power Constructor] [|SMILE API]

Project Status:
This project is working as standalone application and use craigslist website which provides local classified sell purchase of old and new products in which information is available in unstructured, ungrammatical and incoherent data sources. This framework currently dealing three domains of information LAPTOP, CELL PHONE, and CAR to provide a user consolidated search result on one page and eliminate the manual browsing. This project is not only search the available information but predict the unavailable information related to a product which user commonly want to know but not available on web page. Currently, searching and prediction module working separately which require automation.

Project Progress:

 * ** Version ** || ** Task Description ** || ** Issues ** || ** Expected Deadline ** || ** Status ** || ** Comments ** ||
 * 1 || Combine extraction and inference modules || difficulty in automatic discretization || 20 Feb 2010 || completed ||  ||
 * 2 || Development of web application || need to explore how to run java on web || end of march || completed ||  ||
 * 3 || Development of new BN model with large data size for laptop, car, cell phones || need to perform preprocessing to discretize the data. || 10 May || Completed ||  ||
 * 4 || Validation of Data || required to done manually 1000 of record for each domain || in summer ||  ||   ||
 * 4 || Specific Searching capability for web Application || need to explore MySql and its connection with JSP || 15 May || Completed ||  ||