딥러닝 LSTMs 기반 자연어 감성분석 모델을 활용한 거주자의 감성분류Use of Deep-Learning based LSTMs Natural Language Sentiment Classification Model for Residents’ Sentiment Analysis
- Other Titles
- Use of Deep-Learning based LSTMs Natural Language Sentiment Classification Model for Residents’ Sentiment Analysis
- Authors
- 이득영; 장선우; 전한종
- Issue Date
- Oct-2019
- Publisher
- 대한건축학회
- Keywords
- 자연어 처리; 감성 분류; 딥러닝; 건물 성능 평가; LSTMs; Natural Language Processing; Sentiment Classification; Deep Learning; Building Performance Evaluation; Long-Short Term Memory Networks(LSTMs)
- Citation
- 대한건축학회 2019년도 추계학술발표대회논문집, v.39, no.2, pp.168 - 171
- Indexed
- OTHER
- Journal Title
- 대한건축학회 2019년도 추계학술발표대회논문집
- Volume
- 39
- Number
- 2
- Start Page
- 168
- End Page
- 171
- URI
- https://scholarworks.bwise.kr/hanyang/handle/2021.sw.hanyang/146947
- Abstract
- In architectural planning and design, user-centric approach is important since multiple users spend lengthy amount of time in their residences. Post-Occupancy Evaluation(POE) is one effective way of analytically evaluating building performance in terms of users. However, POE is limited in specific project and time consuming. Recently, various types of online platforms have come in use which enables users to provide reviews toward their residences. Those reviews are in natural language form, voluntarily written by users and containing various topic regarding residents’ interests which may suggest one possible way of adopting user-centric approach in building design. In this regard, this paper aims to suggest deep-learning based Long-Short Term Memory Networks(LSTMs) model for sentiment analysis toward user left natural language reviews. For natural language processing and sentiment analysis, this study used “KoNLPy” and “Word2vec for data preprocessinng and Google TensorFlow and Keras for model structure. This approach may suggest one way of recognizing residents’ difficulties and interests toward their residences.
- Files in This Item
-
Go to Link
- Appears in
Collections - 서울 공과대학 > 서울 건축학부 > 1. Journal Articles
![qrcode](https://api.qrserver.com/v1/create-qr-code/?size=55x55&data=https://scholarworks.bwise.kr/hanyang/handle/2021.sw.hanyang/146947)
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.