Research on Customization and Personalized Push System of Micro-Movie Users' Preference Content Driven by Big Data

This paper aims to explore how to build an efficient and accurate micro movie recommendation system through advanced data analysis and machine learning technology, so as to meet users' increasing needs for personalized content. Firstly, the user behavior characteristics of the current micro film market are analyzed, and the characteristics of user preference diversity, timeliness and situational relevance are revealed. On this basis, a hybrid recommendation algorithm combining deep neural network and collaborative filtering is proposed, and real-time feedback mechanism is introduced, so that the system can quickly adjust the recommendation strategy according to the instant feedback of users, and improve the user experience. In addition, considering the protection of user privacy, this study also designed the encryption and anonymization processing scheme to ensure the security of user data while providing personalized services. The experimental results show that compared with the traditional recommendation method, the personalized push system has significantly improved the recommendation accuracy rate, recall rate and user satisfaction. This not only provides a new way to optimize the user experience for the micro film platform, but also points out the creative direction for content creators and promotes the healthy development of the micro film industry.