Abstract
his study proposes a sale price estimation model for reused jewelry using information on product attributes and jewelry price trends. Jewelry products have many factors that affect the price, such as the main gemstone, surrounding decorations, and brands, many of which are recorded as categorical data. In this study, we constructed a model based on the entity embedding model. A comparison with the XGboost model, one of the famous machine learning methods, which is known to be useful for categorical data, confirmed some of the effectiveness of the entity embedding model in terms of generalization performance for test data.
https://ror.org/03qvqb743 
