Query-Based Extractive Text Summarization for Sanskrit

摘要

Sanskrit consists of lots of literature available in the form of epics, stories, puranas, Vedas, and many more. Most of the Sanskrit documents have been digitized and made available online. Searching for the required information from the plentiful documents available is a tedious task. Automatic summarization serves the purpose in such situations. Many tools for summarization have been developed for English and foreign languages. The research for such kind of tools in Sanskrit is under exploration. In this paper, we propose three query-based summary generation methods to obtain extractive summary for single document written in Sanskrit. The methods are based on average term frequency-inverse sentence frequency, the VSM (Vector Space Model) and a graph-based technique using PageRank. All the techniques are compared and evaluated on the basis of performance.