Medical service recommendation is as an essential component of eHealthcare systems, and has received widespread attention in recent years. In medical systems, users can send demands to medical server, which then recommends the suitable doctors based on the demands. In the existing medical service recommendation scheme, although users can send the basic demands to get medical service recommendation, users cannot set the attributes of demands that are more concerned according to their own preferences or personalized demands, so as to get the accurate personalized medical service. In addition, due to the sensitivity of the users’ information, guaranteeing the privacy throughout the recommendation process without sacrificing the accuracy is still challenging. In this paper, we propose a privacy-preserving multi-level attribute based medical service recommendation scheme. This work considers multi-level attributes to fully describe users’ demand information, and users’ concerned attributes are considered to achieve personalized medical service recommendation. We design two algorithms to keep user’s demands secret, and recommend doctors in a privacy-preserving way. Detailed analysis proves that the proposed scheme can achieve the desired security prosperities. Performance evaluations through extensive experiments also demonstrate the efficiency of our scheme.

推荐医疗服务是eHealthcare系统的重要组成部分，近年来受到了广泛关注。在医疗系统中，用户可以将需求发送到医疗服务器，然后服务器根据需求推荐合适的医生。在现有的医疗服务推荐方案中，虽然用户可以发送基本需求以获得医疗服务推荐，但是用户无法根据自己的偏好或个性化需求设置更关注的需求属性，从而获得准确的个性化医疗服务。 。另外，由于用户信息的敏感性，在整个推荐过程中确保隐私而不牺牲准确性仍然是具有挑战性的。在本文中，我们提出了一种基于隐私保护的多层次属性的医疗服务推荐方案。这项工作考虑了多级属性来充分描述用户的需求信息，并考虑了用户的相关属性以实现个性化的医疗服务推荐。我们设计了两种算法来保密用户的需求，并以保护隐私的方式推荐医生。详细分析证明，该方案可以达到预期的安全性。通过大量实验进行的性能评估也证明了我们方案的有效性。我们设计了两种算法来保密用户的需求，并以保护隐私的方式推荐医生。详细分析证明，该方案可以达到预期的安全性。通过大量实验进行的性能评估也证明了我们方案的有效性。我们设计了两种算法来保密用户的需求，并以保护隐私的方式推荐医生。详细分析证明，该方案可以达到预期的安全性。通过大量实验进行的性能评估也证明了我们方案的有效性。