Modeling Kansei Index of Product by Machine Learning Using Review Text and Image

Suzuki, Hidemichi; Tobitani, Kensuke; Hashimoto, Sho; Yamada, Atsuhiro; Nagata, Noriko

doi:10.2493/jjspe.85.1143

Cited by 2 publications

(2 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Kansei indexing method [10]- [14] is a method for quantifying the evaluation factors of a domain. In the Kansei indexing method, the factors of impressions caused by the target domain are determined and the evaluation values of each factor possessed by individuals in the domain are calculated.…”

Section: Related Workmentioning

confidence: 99%

“…In the Kansei indexing method, the factors of impressions caused by the target domain are determined and the evaluation values of each factor possessed by individuals in the domain are calculated. The Kansei indexing method includes methods based on psychological experiments and multivariate analysis methods [10]- [12] and machine learning methods [13]- [14] using review data from the Internet. Psychological experiment-based methods have the problem of high human and temporal load, as they require obtaining Kansei evaluation data through multiple experiments and analyzing them in order to construct Kansei indices with high accuracy.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Visual impression estimation system considering attribute information

Taki

Kato

Terada

et al. 2023

Sixteenth International Conference on Quality Control by Artificial Vision

View full text Add to dashboard Cite

Detailed identification of visual impressions of objects by attributes can be leveraged to develop products and improve customer satisfaction. In this study, we propose a method to estimate Kansei (affective) information for each attribute, which is the visual impression received from the image. For each attribute, we created a dataset with Kansei indices. By fine-tuning the created dataset to combine attribute information with the output of ResNet18 which was already trained with ImageNet to output indexes, we confirmed that the correlation coefficients for multiple item ratings were higher than those of a deep learning model without attribute information.

show abstract

Section: Related Workmentioning

confidence: 99%