Point of Interest recommendation for social network using the Internet of Things and deep reinforcement learning

Abstract

<abstract> <p>Point of Interest (POI) recommendation is one of the important means for businesses to fully understand user preferences and meet their personalized needs, laying a solid foundation for the development of e-commerce and social networks. However, traditional social network POI recommendation algorithms suffer from various problems such as low accuracy and low recall. Therefore, a social network POI recommendation algorithm using the Internet of Things (IoT) and deep reinforcement learning (DRL) is proposed. First, the overall framework of the POI recommendation algorithm is designed by integrating IoT technology and DRL algorithm. Second, under the support of this framework, IoT technology is utilized to deeply explore users' personalized preferences for POI recommendation, analyze the internal rules of user check-in behavior and integrate multiple data sources. Finally, a DRL algorithm is used to construct the recommendation model. Multiple data sources are used as input to the model, based on which the check-in probability is calculated to generate the POI recommendation list and complete the design of the social network POI recommendation algorithm. Experimental results show that the accuracy of the proposed algorithm for social network POI recommendation has a maximum value of 98%, the maximum recall is 97% and the root mean square error is low. The recommendation time is short, and the maximum recommendation quality is 0.92, indicating that the recommendation effect of the proposed algorithm is better. By applying this method to the e-commerce field, businesses can fully utilize POI recommendation to recommend products and services that are suitable for users, thus promoting the development of the social economy.</p> </abstract>

Joint Urban Modeling With Graph Convolutional Networks and Crowdsourced Data: A Novel Approach

“…To facilitate end-to-end modeling of urban data, POIs emerge as a highly valuable data source. POIs are intrinsically linked to group behavior [12] and the socio-economic facets [13] of urban areas, making them one of the most widely utilized sources of crowdsourced data currently . Furthermore, POIs are more readily available compared to other data sources (e.g., group mobility data), which are typically restricted to specific regions and user groups.…”