Paper Abstract and Keywords |
Presentation |
2019-09-27 17:50
Improvement of Comment Ranking in Yahoo! News by In-House Competition Hiroaki Taguchi (Yahoo Japan), Soichiro Fujita (Tokyo Tech), Hayato Kobayashi (Yahoo Japan/RIKEN), Yoshimune Tabuchi, Ken Kobayashi (Yahoo Japan), Kazuma Murao (VISITS Technologies), Chahine Koleejan, Takeshi Masuyama, Taichi Yatsuka (Yahoo Japan), Manabu Okumura (Tokyo Tech) NLC2019-16 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
In this paper, we report on efforts of an in-house competition held for improving models that rank user comments in Yahoo! News in terms of their constructiveness.
Specifically, we describe
(1) settings of the competition such as task and evaluation,
(2) behavior of participants and submitted models,
(3) performance comparison of the submitted models and the current model,
(4) performance improvement by an ensemble of the submitted models,
and (5) future issues based on the questionnaire answers of the participants. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
News Comment Ranking / Model Ensemble / In-House Competition / / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 119, no. 212, NLC2019-16, pp. 41-46, Sept. 2019. |
Paper # |
NLC2019-16 |
Date of Issue |
2019-09-20 (NLC) |
ISSN |
Online edition: ISSN 2432-6380 |
Copyright and reproduction |
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
Download PDF |
NLC2019-16 |
|