Research Of The Service Pushed Initiatively Method Based On Cloud Manufacturing Service Platform

The rapid development of the Internet Economy has brought new vitality to the industry of Manufacturing, Cloud Manufacturing Service Platform was born at the right moment, in which contained advanced technologies such as Cloud Computing, the Internet, Virtualization, Cloud Manufacturing and so on, and it is widely applied in the enterprise rapidly. The demand side and the service provider from all over the country can find a satisfied partner quickly through the Platform, which can extremely improve degree of needs, services, and resources sharing. Users finding partners becomes more and more difficult with the increasing number of users, demands, services, resources in the Platform. Search by inputting keywords manually takes user’s too much time, and the search results can not always keep users satisfied every time. So the research of push service initiatively for users has become a hot spot nowadays.IBM’s "smarter planet" concept provides new opportunities for the transformation and upgrading of Manufacturing Industry, then the Wisdom Cloud Manufacturing techniques appears. This article is putting the wisdom concept into Cloud Manufacturing Service Platform to put forward the concept of service push initiatively, and pushing service initiatively for users aims at solve the problem that users cannot find partners quickly and accurately. Aiming at the problems existing in the Platform now, first of all, we described the model of user requirement in the Platform, and Proposed the Service Pushed Initiatively Method Based on the Platform. This method is a combination of multilevel hybrid push methods, which including the Clustering Analysis Method, Collaborative Filtering Method, Service Evaluation Method Based on User Preferences. Through the combination of explicit requirements and implicit preferences, this method has avoided the data sparse, cold start issues of Collaborative Filtering Method to a certain extent, and improved the accuracy of the services pushed. Secondly, we elaborated the key technologies of the Service Pushed initiatively Method in detail, including mine the user’s demands, build user’s preference model, and design the service push algorithm. We use Web mining method acquiring implicit demand, and use the Vector Space Model describing preference Model through TF- IDF function to calculate the proportion of user’s interests and preferences, and use Cosine Angle formula, Person formula calculating similarity. Finally, we verified the Service Pushed Initiatively Method through a researching project in laboratory.