Cloud Manufacturing Service Recommeendation Based Of Improved Spectral Clustering And Slope One Algorithm

Cloud manufacturing service providers publish manufacturing resources or business functions to the cloud manufacturing platform through virtualization and servitization.Users can search the cloud manufacturing services they need in the cloud platform.By renting these cloud manufacturing services,they can realize the business beyond their own manufacturing capabilities and quickly build value-added manufacturing services.There are many process attribute parameters in cloud manufacturing services.These attribute parameters have rich types and large value space difference.It is a challenge to recommend cloud manufacturing services.In addition,users’ historical service usage habits,service preferences and the popularity of cloud manufacturing services all have an important impact on cloud manufacturing service recommendation,which further increases the difficulty of cloud manufacturing service recommendation.Therefore,a cloud manufacturing service recommendation method based on improved spectral clustering and Slope One algorithm is proposed in this paper.This method introduces service clustering to improve the service discovery efficiency in the recommendation process,and corrects the service deviation in Slope One algorithm to enhance the rationality and recommendation accuracy of cloud manufacturing scores.The main work and contributions of the paper are as follows:(1)Cloud manufacturing services are crawled from multiple cloud platforms,and a simulation validation dataset of cloud manufacturing service recommendation with service scores is constructed.The proposed clustering method is superior to other spectral clustering algorithms in terms of clustering indicators SC,CH,NMI and FMI.The proposed method is superior to the current popular methods in terms of service scoring and recommendation accuracy.(2)A spectral clustering algorithm integrating multi-dimensional similarity for cloud manufacturing services is constructed.The similarity calculation methods of textual and numerical attributes are presented,respectively.An attribute similarity fusion function is designed to realize the fusion of multi-dimensional cloud manufacturing service attribute similarity.The cloud manufacturing service similarity matrix is constructed based on cloud manufacturing service attribute similarity.The intrinsic gap is introduced to determine the number of clusters in the cloud manufacturing service spectral clustering.By the introduction of service clustering,the service search space in the recommendation process is reduced,and the candidate scoring service set is efficiently constructed.(3)An improved Slope One algorithm for cloud manufacturing service recommendation is proposed.The user similarity is calculated by considering three factors: service rating,rating time weight and service popularity.The user similarity and service similarity are introduced to improve the quality of service deviation calculation,and the service deviation weights corresponding to different cloud manufacturing services are also modified.It can improve the accuracy of Slope One algorithm score prediction and recommendation.