A statistical service composition approach based on Hidden Hierarchy Markov Model

How to composite Web service is one the most important problems of next generation Internet (NGI) researching. In order to reuse complex service for service composition, multi-scale Web service composition (MSWC) is proposed. Inspired by word segmenting algorithm, service composition can be achieved by diving service request into several groups and mapping these groups into existed complex services. The novelty of this approach is describe service composition domain as a hidden hierarchy Markov model, and service segmenting and mapping are based on service concurrent probability and logic probability respectively. MSWC is especially more effective, because if goal service already exists, this approach is transformed into a search method.