A carregar...
Publicação
Machine learning based art price forecasting: generalization for selected Belgian painters
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 733.5 KB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
This study uses Machine Learning (ML) to predict prices for paintings by 11 selected Belgian
artists based on data from Christie’s spanning the period from 1994 to 2023. Examining the
nuanced relationship between representation and reality, the research reveals ML's limited
accuracy compared to human experts. While the Random Forest model outperforms Gradient
Boosting, constraints such as a small dataset impact generalization. Recommendations include
expanding the dataset and integrating features related to online information and the number of
bidders. Despite current disparities, ML shows promise as a complementary tool to enhance
efficiency in the appraisal landscape of paintings.
Descrição
Palavras-chave
Finance Forecasting Machine learning Art Painting Belgium Regression Random forest Gradient boosting Ensemble methods Web scraping