Abstract:

With the development of the service technology, more and more organizations publish their business function as service through Internet. With the quick increase of service number, the cost of classifying these services manually is becoming more and more expensive. An automatic service classification method based on WordNet is proposed by combining text mining, semantic technique and machine learning technique. In this approach, text mining and semantic disambiguation techniques are used to obtain a set of Sense vectors with exact word meaning from description documents and social annotation of the service. These Sense vectors can describe a service. They are a set of tags of socially annotating each API. Then, a Kmeans algorithm is used to classify these services. At last, using the services on Programmable Web as test data, experiments demonstrate that our classification method has good effect.

Key words: service;word sense disambiguation;social tagging;similarity;WordNet