Words of scents: A linguistic analysis of online perfume reviews
This study explored linguistic resources that people employ to express their perceptions and opinions of a fragrance. Several natural language processing (NLP) techniques were used, including sentiment analysis, topic modelling, and supervised classification. The data were collected from the website...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Published: |
Walter de Gruyter GMBH
2024
|
| Subjects: | |
| Online Access: | http://eprints.um.edu.my/44328/ https://doi.org/10.1515/lingvan-2022-0135 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Universiti Malaya |
| Summary: | This study explored linguistic resources that people employ to express their perceptions and opinions of a fragrance. Several natural language processing (NLP) techniques were used, including sentiment analysis, topic modelling, and supervised classification. The data were collected from the website of Fragrantica, popular among perfume lovers, and the reviews pertained to a niche market fragrance. The findings from the sentiment analysis revealed that the positive reviews contained a greater percentage of lexis referring to the social world, social processes, sexuality, leisure, and mental states. In the negative reviews, the prevalent lexis concerned health, illness, and social conflict. The topic modelling analysis indicated that the positive reviews centred on issues related to the experience of wearing the perfume, its sensual character, and its olfactory qualities. The negative reviews focused on skin chemistry, the perfume notes, and physical reactions to the scent. The supervised classification analysis indicated notable differences in the frequency ratios of some lexical items in the positive and negative reviews. |
|---|