[9p-W321-1~14] 23.1 Joint Session N "Informatics"

△ [9p-W321-5] A study on classification of meeting abstracts using Doc2Vec

〇(DC)Kohei Ishikawa¹, Kentaro Kutsukake², Shunta Harada^1,3, Miho Tagawa^1,3, Toru Ujihara^1,3,4 (1.Nagoya Univ., 2.RIKEN AIP, 3.IMaSS, 4.GaN-OIL AIST)

Keywords:Natural language processing, Data science, Data mining

There are many oral or poster presentations held in an academic conference. Generally, meeting abstracts are prepared for participants to search presentation. However, It is impossible for them to look through all text in abstracts. Consequently, participants are forced to search abstracts by their title, although the title sometimes represents the contents inappropriately. To address this issue, we investigate the feasibility of auto classification of meeting abstracts using machine learning. In this research, we used Doc2Vec to obtain the distributed representation of abstracts.