LAUSR

With the advent of sixth generation (6G) communication systems, edge computing benefits from enhanced bandwidth and increased network throughput, further diminishing service latency. Against this background, 6G-enabled edge computing makes it possible for service recommendation to reach more precise user profiling and efficient personalization. However, existing service recommendation algorithms are limited to performing coarse-grained service preference prediction based on historical data within specific time periods, which is incompatible with high frequency of information updating of 6G networks. Fortunately, digital twins (DTs) can model complex service recommendation systems and facilitate prediction of user service preferences in real-time with simultaneous data interaction. To this end, a DTs-assisted service recommendation scheme with preference prediction in 6G-enabled edge computing, named DPSR, is proposed. Specifically, DPSR utilizes DTs to perform modeling of service recommendation problems and to continuously obtain user data in real-time. Subsequently, a lightweight time-series prediction model is deployed on the DTs to predict user service preferences, thereby identifying users with a higher likelihood of requesting services in the next moment. Finally, a collaborative filtering (CF) model is used to provide personalized service recommendations for these users. The effectiveness of the DPSR is demonstrated in the experimental section of this article.