Abstract: In this era of information explosion, personalized recommendations have come into being. Based on user preferences, personalized recommendation systems can actively recommend products or information that they may be interested in. This can not only free users from complicated products in e-commerce, but also increase sales. Most of the existing researches are concentrated more on the field of information science and computer science to improve the accuracy of the algorithm. Some scholars in marketing, statistics and other fields pay more attention to the effect of personalized recommendation system on user behavior and user satisfaction. Most of them use methods such as theoretical analysis and questionnaires. As the user experience and feelings are more abstract, the closed questions of the questionnaire survey method are difficult to obtain the true feelings of users. Taobao platform is selected to conduct empirical researches and situational interview methods are used to obtain user feelings about the recommendation system. Meanwhile, the indicators that users pay more attention to are obtained such as perceived usefulness, perceived security, perceived comfort, and ease of use. Finally, some suggestions for the recommendation system of e-commerce platforms are put forward.

Key words: personalized recommendation, user perception, e-commerce