Abstract: The rise and development of mass customization models have effectively alleviated the contradiction between users' demand for differentiated and personalized products and the high cost of pursuing customized products. In order to more efficiently assist users to make satisfactory product customization decisions in the process of mass customization, the traditional mass customization-oriented recommendation algorithm is improved accordingly, and combined with the characteristics of mass customization, a recommendation scheme based on user portraits is proposed. The item-based collaborative filtering algorithm is selected as the basic recommendation algorithm, and the big data tool user profile model is introduced to filter the initial recommendation results twice, so as to improve the traditional collaborative filtering recommendation algorithm to ignore the user's own interest preference and improve the user customization experience and recommendation accuracy. The case of mobile phone product customization is given, and the whole process of generating the final recommendation result is simulated and analyzed, which verifies the effectiveness and feasibility of the recommendation algorithm.

Key words: mass customization, collaborative filtering, user portrait, intelligent recommendation, big data