Deep learning-based natural language sentiment classification model for recognizing users' sentiments toward residential space

Abstract

Recent developments in real estate brokerage platforms have enabled residents to provide subjective reviews, which have immense value as subjective assessments and suggestions for architects. This study suggests a deep-learning-based natural language sentiment classification model to analyse reviews. Morpheme analysis and word embedding for 'KoNLPy' and 'Word2vec' were structured for pre-processing, and a long short-term memory network was used to process review data. Total 5974 review data were used in this study. Among the various active online platforms for real estate brokerage, platforms that provide online users with the ability to write reviews of their living spaces were crawled. The review data were classified as 'positive' or 'negative' by label and as 'Apartment' or 'Non-Apartment' by housing type. The model developed in this study is expected to increase in value as more online platforms appear in the future and the volume of natural language data generated by those platforms increases.