Analysis of recommendation algorithms for Internet of Things

摘要

Internet of Things (IoT) is a new paradigm that refers to a world-wide network of interconnected physical things using standardized communication protocols to provide human useful services such as personal health care and green energy monitoring. Nowadays, many third party service providers are providing a large number of IoT services. Recommending IoT services to users based on objects they own will become very crucial for the success of IoT. In this paper, we investigate the possibilities of leveraging recommendation algorithms, especially graph-based, to IoT. We propose a hyper-graph model for IoT systems in which each hyper-edge connects users, objects, and services. Next, we conduct experiment to analyze and explore correlations between performances of different algorithms on IoT Service Recommendation (IoTSRS) based on existing well known metrics. Results show that the graph-based recommendation algorithm can be used to develop an effective recommender system for IoT. Moreover, we show that some algorithms perform reasonably well and produce high quality results. However, further extension of existing approaches is required for IoTSRS.