Cited 0 time in webofscience Cited 0 time in scopus

전문분야 문서 분류를 위한 불균형 데이터 처리 방법

Title
전문분야 문서 분류를 위한 불균형 데이터 처리 방법
Translated Title
Deep Learning based Imbalanced Data Processing Methods for Special Documents Classification
Authors
진상현강원석손창식
DGIST Authors
진상현; 강원석
Issue Date
2019-11-16
Citation
2109 대한임베디드공학회 추계학술대회, 298-300
Type
Conference
ISBN
9788996655312
Abstract
In this paper, we propose a document classifier using natural language processing and deep learning to develop a training system and propose a method for improving accuracy through class imbalanced data processing. To train the documents classification model, the aircraft maintenance documents were preprocessed using KoNLPy and extracted features based on the TF-ICF. In addition, the documents were classified through the deep learning-based classifier consisted of four convolutional layers and two fully-connected layers. As a result, the accuracy of classifying documents was improved by 2.2% on macro-average using the proposed imbalanced data processing method. Especially, the document classification accuracy increased by 6% in the minority class.
URI
http://hdl.handle.net/20.500.11750/14142
Publisher
대한임베디드공학회
Related Researcher
  • Author Kang, Won-Seok  
  • Research Interests Data Mining & Machine Learning for Text & Multimedia, Brain-Sense-ICTConvergence Computing, Computational Olfaction Measurement, Simulation&Modeling
Files:
There are no files associated with this item.
Collection:
ETC2. Conference Papers


qrcode mendeley

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

BROWSE