A Fusion of Granulation and Artificial Neural Network: A New Service Selection Method
Huiling Wang, Bin Zhu, Guanyu Li, Piyun Teng · 2016
Since similar-functioned web services are gradually increasing in Semantic Web of Things, it's difficult for users to select services they need quickly and efficiently. To solve this problem, this paper proposes a mono-service oriented selection method - Granular Computing Neural Network Service Selection (GCNNSS), which combines Granular Computing with Back-propagation Artificial Neural Network (BPANN). Firstly, match users' service request description with services in services repository to generate satisfying candidate services set. Then, filter the original attribute set of candidate services by service graining, using the minimum attribute set algorithm to get minimum attribute set of original attribute set. Finally, reorder the candidate service set by using the iterative learning method of BPANN and return the result to users. According to the contrast research and simulation experiment, the proposed method GCNNSS is superior to other representative service selection methods in effectiveness and reliability.