摘要：A social network site is one social structure made up of a set of users including individuals or organizations, which plays a very import pole in the digital age. It has been one type of fashion online platform at providing services on facilitating the establishment of social networks or social relations among social members. Most social network sites provide web-based services and web-based means that allow users to interact over the internet to share individual experiences and spread information. Thus, the user-online load movement analysis is increasingly important for one social network site because of its significant effect on resource allocation, web traffic, maintenance management and economy of operations. Among the varying soft computational tools and algorithmic models available, the back-propagation artificial neural network (BP-ANN) model is one of the most commonly used and robust models. In this study, a typical BP-ANN with a single hidden layer is employed for forecasting the user-online load movement. Experimental results of the user-online load movement forecast at several social network sites show workability the proposed method.
