Today I met my supervisor, Dr Ruvan Weerasinghe and we discussed the possible paths to carry out my research. We discussed the possibilities of implementing Banko’s approach which he described on his paper, “Headline Generation Based on Statistical Translation” (Banko, M., Mittal, V.O., and Witbrock, M.J. 2000).
Then due to lack of pre summarized text for the training, we decided to model an extractor, which can extract summaries from a source. It was decided to use basic surface level approaches to rank the sentences and weight them. Then the summary will be evaluated using the question answering approach. So first I need to find set of data and made questions from those by humans. Then the summary can be evaluated using that data.
Dr Ruvan gave me the book, “Text Mining Application Programming” by Manu Konchady and I read two chapters of it which explained the methods of information extraction and Summarization.
Tuesday, September 18, 2007
Subscribe to:
Posts (Atom)