S.-I. Kazuhiro (Japan)
important sentences, natural language, morphological anal ysis, extracting algorithm, given problem
In this paper, we propose an algorithm for extracting im portant sentences from papers which students submit for solving problems. Such a sentence is defined by not only high frequency words but also words with relation to key words calculated by using this algorithm. The evaluation value of each sentence is calculated on the basis of each evaluation value of words in the sentences. The experimen tal results show that the correlation coefficient between the priority selected by the human markers and that calculated by using this algorithm is high. As this algorithm in this paper is for writing in English, it should be very useful to analyze the content of the papers submitted by students in a short time.
Important Links:
Go Back
