Document summarization apparatus, document summarization method, and recording medium recording document summarization program

摘要

A document summarizing apparatus generates a comprehensive summary on a group of documents of relatively diverse contents. The structure of documents specified to be processed is analyzed in a phrase analyzing unit to generate analytic trees describing the dependencies between words. An analytic tree scoring unit adds scores to the analytic trees in accordance with their importance. An analytic tree score accumulating unit accumulates scored trees to unify the trees expressing the same concept to increases the scores added to the unified analytic trees. A sentence synthesizing unit then selects the trees with higher scores from within the set of analytic trees stored in the analytic tree score accumulating unit to synthesize a summary from the selected analytic trees. The present invention allows less limitation to be applied to the documents to be processed, as well as a comprehensive summary to be generated.