This paper investigates a traditional cultural resource recommendation method integrating knowledge graphs and interest preferences. Based on this approach, a new media cultural resource dissemination platform with functions including traditional cultural resource recommendation and user search has been designed. Taking a deep learning-supported new media cultural dissemination platform as the research subject, this study employs questionnaire surveys and mathematical statistical methods to conduct a sample survey across six universities in City A. It quantitatively evaluates the platform’s specific impacts on two levels: educational effectiveness and user experience for audiences of traditional cultural education. Findings indicate that the platform’s implementation increases audiences’ willingness to engage with traditional culture and elevates the proportion of positive emotional responses toward traditional culture. The number of individuals familiar with traditional culture increased by 45.7%, significantly enhancing teachers’ and students’ understanding of traditional culture, spiritual and cultural cultivation, participation levels, and satisfaction with cultural dissemination. User experience scores across 14 dimension indicators averaged no less than 3.5 points. This shows that the new media cultural dissemination platforms based on deep learning are a good channel through which traditional culture can find its way into the world, which has a positive effect on traditional culture education.
