| In the field of logistics, improved enterprise credit monitoring helps toimprove the image and competitiveness of logistics enterprise. This paperstudies the related technologies of logistics enterprise credit rating in thebackground of a provincial transportation logistics cloud platform project.Currently logistics credit rating mainly meets the following two questions.1) The main methods of logistics enterprise credit rating are empiricalmethods, which are not accurate and objective enough.2) Data of logisticsenterprise such as law enforcement data and transportation data is stored inthe government systems such as comprehensive enforcement system andtransport information management system, which is related to privacy. Weneed to do data privacy preserving when integrating is running.In response to above problems, this paper proposes an improvedprivacy preserving data integration approach based on differential privacyand a logistics enterprise credit rating method based on Bayesian network.Then we design and implement a system of logistics enterprise credit ratingusing those two methods. The feasibility and effectiveness of the system isproven by the data validation.Compared with other similar systems, this paper works with thefollowing characteristics:(1) Propose a privacy preserving data integration approach usingdifferential privacy. This approach implements distributed noise generationto solve the problem that noise is unconctrollable in data integration andimplements anonymous multi-set operation to solve the problem that Joinoperation exposes critical data to other data sources..(2) Propose a logistics enterprise credit rating method based on Bayesian network. This method provides improved convergence time andaccuracy of Bayesian network structure to improve the accuracy of logisticsenterprise credit rating. Compared with ACO-B algorithm, K2score is up to1.75and convergence time is5.87%faster.(3) Based on the above work, this paper completes the architecturedesign and implementation of a logistics enterprise credit rating system,which includes private data integration plan client, private data integrationengine and credit rating engine.(4) Apply the system to the transportation logistics cloud platformof one province and provide SaaS service of credit information for the cloudplatform user. Trail run shows that this system has98.50%rating correctrate. |
