Technology For User's Information Push Based On Machine Learning In Cloud Manufacturing Environment

The birth of the Internet has brought a new life to the manufacturing sector,the integration of cloud computing,networking,virtualization and other advanced cloud manufacturing service platform are created.Cloud manufacturing as a new service-oriented manufacturing model makes China's manufacturing industry more powerful,and transforms it from primary manufacturing to service-oriented manufacturing.Therefore,through the assistance of the cloud manufacturing service platform,the service demander and service providers around the world can find their satisfactory resources conveniently and efficiently.Nowadays,the increasing information in the cloud manufacturing service platform makes users spend a lot of time and energy to refine the effective information from the different kinds of the information,which seriously hindered the service demander for seeking the suitable partners and made the service providers recommend related services to service demander more difficultly.Thus,it has become a hot research topic for the service provider to actively push forward the effective service information.This paper studies the technology of the user's information push by the meanings of the Machine Learning in the cloud manufacturing environment.In the aspect of the relevant theoretical basis,this paper firstly describes the background and the value of cloud manufacturing,and introduces the corresponding research and the application of cloud manufacturing.Then,it introduces the theoretical knowledge of social manufacturing,the concept of machine learning technology and the classification of recommendation system.Finally,it summarizes the deficiencies and the value of the user's information push technology in the current cloud manufacturing environment,and the construction of user's information push system under the background of social manufacturing by the use of machine learning are presented.In the aspect of the data acquisition and processing technology,the software of web crawler tools is developed at first,which accesses to relevant information about the theme and unauthorized users from cloud manufacturing platform by use of the open API interface.After the completion of the data collection,through the operation of text segmentation stop,word processing and the selection of interest feature,the standard user's interest feature vector and each user's interest feature vector are established.Then,the user's interest feature vector used in the machine learning is seen as the input parameter,namely the establishment of user clustering method based on multi-Markov model,which classifies the users having different interest and clusters the users having the same interest.Finally,the users having the same interest utilize the collaborative filtering method based on the user's interest,and the list of user's love can be achieved by using the model of "user-service" evaluation.Then,the user's information push can be completed offline.In the aspect of the system implementation and experimental comparison,we firstly obtain the relevant information on "Chinese machinery manufacturing platform",and then preprocess the obtained experimental data.The treated experimental data is put into the construction of the push model to realize the test of the service push,which verifies the validation of the construction of the push system by the machine learning in cloud manufacturing environment.Secondly,we find that the user's clustering effect will influence the results of the experimental comparison before collaborative filtering operation.If the better pre-clustering effect is,the more accurate the latter service push is.Therefore,in the following study,it is necessary to establish a more perfect model of the "user-service" evaluation,as well as a more efficient and accurate method of the user clustering in order to improve the accuracy of the technology of the user's service push.