Application of a Personalized Recommendation Method for Public Art Resources Based on Regional Cultural Attribute Extraction in Archaeological Sites

Abstract

The regional cultural attribute significantly influences users' perception of artistic heritage, making it challenging to ensure the hit rate of public art resource recommendations solely from a behavioral perspective. Therefore, a personalized recommendation method for public art resources based on the extraction of regional cultural attributes is proposed, with its application in archaeological site resource recommendations. By combining the regional category and type of cultural attributes, the regional cultural attributes are extracted through classification. Subsequently, the Q-learning algorithm is employed to fit and calculate the regional cultural attributes of the target recommendation object and the cultural attributes of the resources in the public art resource pool. A loss function is introduced to filter the fitting results, and finally, the TOP-N in the public art resource sequence with a fitting relationship is taken as the recommendation result. In the test results, the hit rate of the recommended outcomes consistently remains above 0.7, with a maximum value of 0.96. When the parameter α increases from 3 to 15, the corresponding hit rate improves by 0.26. The application of this method in archaeological site resource recommendations not only enhances resource utilization but also provides a new perspective for archaeological research and cultural heritage preservation.