Trích rút quan hệ giữa các thực thể từ văn bản Tiếng Việt sử dụng phương pháp lan truyền nhãn

Abstract. This paper presents a relation extraction system for Vietnamese texts using label propagation. In this paper, we propose: (i) a measure of similarities between two sentences; (ii) a method

to decrease the effect of high frequency labels in the documents. Our experimental results show that

proposed label propagation method achieves a higher accuracy than the ordinary one [10]. Moreover,

its accuracy is also higher than the support vector machine method applied.