With the rapid development of Internet of Things, mobile edge computing which provides physical resources closer to end users has gained considerable popularity in academic and industrial field. As the number of edge server increases, accessing effective edge services fast is an urgent problem to be solved. In this paper, we mainly focus on the cold-start problem for service recommendation based on location of users and services. Address this conundrum, we propose a service recommendation method based on collaborative filtering (CF) and location, by comprehensively considering the characteristic of services at the edge, mobility and demands of users at different time periods. In detail, we synthesize the service characteristics of each dimension in different time slices through multidimensional weighting method at first. Then We further introduce the idea of Inverse CF Rec to the traditional CF and predict the lost quality of service (QoS) to solve the problem of sparse data. Finally, a recommendation algorithm based on predicted QoS and user geographic location is proposed to recommend appropriate services to users. The experimental results show that our multidimensional inverse similarity recommendation algorithm based on time-aware collaborative filtering (MDITCF) outperforms Inverse CF Rec in terms of the accuracy of recommendation.

随着物联网的飞速发展，提供物理资源更接近最终用户的移动边缘计算已在学术和工业领域获得了极大的普及。随着边缘服务器数量的增加，快速访问有效边缘服务是一个亟待解决的问题。在本文中，我们主要针对基于用户和服务位置的服务推荐冷启动问题。针对这一难题，我们通过综合考虑边缘服务的特性，移动性和不同时间段的用户需求，提出了一种基于协同过滤（CF）和位置的服务推荐方法。详细地，我们首先通过多维加权方法来合成不同时间片中每个维度的服务特征。然后，我们将反向CF Rec的思想引入传统CF中，并预测服务质量损失（QoS），以解决数据稀疏的问题。最后，提出了一种基于预测的QoS和用户地理位置的推荐算法，以向用户推荐合适的服务。实验结果表明，基于时间感知协同过滤（MDITCF）的多维逆相似推荐算法在推荐精度方面优于逆CF Rec。