基于WEB的服务发现流研究

通过搜集和分析WEB用户行为数据，提出了研究WEB用户行为的一个新的视角，即把人看作是传播的内容，把信息资源看作是对象，在这种思路下，WEB网络可以被看作是一个人类集体服务发现流在信息资源之间分配和流动的网络，即服务发现流网络．应用向量自回归条件异方差模型设计了服务发现流预测算法，为了管理不确定性，引入了一些随机变量，这些随机变量会不断地被预测的各个阶段所积累，但通过使用时间序列分析大大地减轻了这种状况．

Research on service discovery stream based on WEB

By collecting and analyzing the data on the behaviors of WEB users, in this paper, the author proposes a new perspective on researching the behaviors of WEB users, i.e., WEB users are regarded as content propagated in WEB and the information resources are regarded as object. By this means, the WEB may be regarded as a network that a humanity collective service discovery stream is assigned and flows among information resources, i.e., the service discovery stream network. The author applies the vector autoregressive conditional heteroscedasticity model （VARCH） to the design on the prediction algorithm of the service discovery stream. In order to manage the uncer,tainty, some random variables are introduced, which may be accumulated continually at each prediction stage. However, this situation is mitigated greatly by the application of the analysis on time series.