Cultural Product Valuation Based on XGBoost-LightGBM

With the prosperity of the cultural industry, it has already become a pillar industry of the national economy. In contradiction to the development trend of the cultural industry, the value determination and valuation of cultural products are still in a state of absence, which restricts the development of the trading market of cultural products. In the traditional way, cultural products are appraised by appraisal experts. With the popularity of the Internet and the development of digital technology, the appraisal method of cultural products has changed, and the study of value appraisal has become an important topic in the whole field of cultural research. In this paper, the possibility of using a machine learning algorithmic framework to evaluate the value of calligraphy and painting works is investigated. Customized painting and calligraphy-related datasets in unstructured data format are used as the main form, and a machine learning-based XGBoost-LightGBM combination method is proposed for evaluating the value of cultural products. In addition, the structured and unstructured data are pre-processed, and the long text is transformed into vector form using the Bert-based model. In addition, the structured and unstructured data were preprocessed, and the long text was transformed into vector form using the bert-Chinese model, and the results were finally input into the XGBoost and LightGBM weighted combination models to obtain the final estimated value range. The experimental results further show that the XGBoost-LightGBM combination method based on machine learning proposed in this paper can be applied to the field of cultural product appraisal.