In this study, the characteristics of the short platform content platform personalization recommendation system were identified, and the effect of the characteristics on perceived value and intention to continue using was identified. In addition, the ...
In this study, the characteristics of the short platform content platform personalization recommendation system were identified, and the effect of the characteristics on perceived value and intention to continue using was identified. In addition, the mediating effect of perceived value on the effect of continuous use was empirically analyzed and verified that the perceived risk level of users had a moderating effect between the characteristics of the personalized recommendation system and the perceived value. Chinese users with experience using TikTok's personalized recommendation system were selected as the subject of the survey. The survey period was conducted for five days from October 29, 2023 to November 2, 2023, and a total of 550 questionnaires were collected by distributing the produced questionnaires online. A total of 518 questionnaires were used for empirical analysis, excluding those with no experience or unfaithful answers in the TikTok personalization recommendation system.
The collected data were analyzed using statistical programs SPSS 26.0 and AMOS 23.0. Depending on the purpose of the study, the moderating and mediating effects were verified by frequency analysis, exploratory factor analysis, reliability analysis, path analysis, hypothesis verification, and multiple regression analysis.
In summary, it was found that the accuracy, novelty, popularity, and ease of the personalization recommendation system had a significant positive (+) correlation with the user's intention to continue using through the full mediation effect of the perceived value, and the privacy exposure of the personalization recommendation system had a significant negative (-) correlation with the user's intention to continue using through the full mediation effect of the perceived value. It was found that the reliability of the personalized recommendation system had a positive (+) correlation with the user's intention to continue using through the partial mediating effect of the perceived value, and the perceived usefulness of the personalized recommendation system had a significant positive (+) correlation with the user's intention to continue using. It was found that the diversity and perceived personalization of the personalization recommendation system did not significantly correlate with the user's intention to continue using. The relationship between the accuracy, diversity, popularity, reliability, ease, perceived usefulness, perceived personalization, privacy exposure and perceived value of the personalization recommendation system was found to be controlled by the perceived risk level, and the novelty of the personalization recommendation system had no moderating effect.
Based on the results of this study, it was suggested that it should be important to develop a personalized recommendation system for a short platform content platform, and it is expected to be meaningful basic data for companies to contribute to the development of future services from the perspective of users' privacy exposure.