With the development of Internet, the problem of information overload becomes more and more serious. The personalized recommendation technology can establish user files through the user’s behavior and other information, and automatically recommend the items that best match the user’s preferences, thus effectively reducing the information overload problem. Based on this, this paper studies the personalized recommendation algorithm based on user preferences in mobile e-commerce. In this paper, user preference model under UTA algorithm is constructed on the basis of user rating on multiple criteria of the project, and user preference clustering is used to improve the scalability problem of personalized recommendation. Finally, the simulation is conducted according to the proposed personalized recommendation algorithm based on user preference. The simulation data use the multi-criteria rating data from 6078 users of Yahoo! Movies website for 976 movies (including 62,156 rows of data). The experimental results show that the multi-criteria recommendation algorithm (MC-CF-dis), which uses user distance similarity, has the best effect, and the MAE and RMSE value of this algorithm is about 1.2 lower than that of the other three algorithms. Accuracy is 6–10% higher than other algorithms. Thus, using this personalized recommendation algorithm based on user preference can effectively improve the quality of recommendation.

随着互联网的发展，信息过载的问题越来越严重。个性化推荐技术可以通过用户的行为和其他信息建立用户文件，并自动推荐与用户的偏好最匹配的项目，从而有效地减少了信息过载的问题。基于此，本文研究了移动电子商务中基于用户偏好的个性化推荐算法。本文基于用户对多个项目标准的评价，建立了基于UTA算法的用户偏好模型，并通过用户偏好聚类来解决个性化推荐的可扩展性问题。最后，根据所提出的基于用户偏好的个性化推荐算法进行了仿真。模拟数据使用来自Yahoo! 6078用户的多标准评分数据。电影网站上有976部电影（包括62,156行数据）。实验结果表明，利用用户距离相似度的多准则推荐算法（MC-CF-dis）效果最好，该算法的MAE和RMSE值比其他三种算法低约1.2。 。准确性比其他算法高6–10％。因此，使用基于用户偏好的个性化推荐算法可以有效地提高推荐质量。效果最好，该算法的MAE和RMSE值比其他三种算法低约1.2。准确性比其他算法高6–10％。因此，使用基于用户偏好的个性化推荐算法可以有效地提高推荐质量。效果最好，该算法的MAE和RMSE值比其他三种算法低约1.2。准确性比其他算法高6–10％。因此，使用基于用户偏好的个性化推荐算法可以有效地提高推荐质量。