Google Pagerank based text summarizer

Members:

3

Phase I In essence we convert the document to a forward graph , each node is a sentence. Edges are labelled with weights (Weights are calculated based on vector space model which amounts to similarity) . Phase II Then we run Google pagerank algorithm iteratively (typically 5 iterations were enough for convergence) which assigns scores to nodes (i.e sentences). We sort the sentences and choose the top 'n' sentences as the summary.