Research On Collaborative Regression Joint Learning Algorithm Of Coupled Privacy Features

In recent years,with the booming development of big data industry,protecting data privacy and security has become a global hot issue.For example,Facebook suffered a data breach in 2018 that wiped more than $36 billion off its market value.In order to solve the problems of data privacy and security,Google proposes a joint learning framework which enables effective and accurate use of data across different organizations.Suppose two organizations,A and B,each maintaining a private record of different sets of features for a common entity.If there is no coupling between the features,the processing is relatively simple.Assuming strong coupling between features,how to deal with it is relatively complicated,both to ensure privacy and to obtain coupling relationships,such as inner product calculation results.In order to solve the above problems,we propose and implement an attribute regression framework based on joint learning.By organically combining joint learning and regression analysis,the framework can realize the coupling relationship between multiple nodes to obtain data efficiently on the premise of protecting data privacy.The experimental results show that the scheme is feasible,data privacy is guaranteed,the feature coupling between different organizations is calculated,the result is accurate,and the execution time is acceptable.Due to the need to transfer a large amount of processed data between different nodes and the limitation of network bandwidth,the implementation time of this framework is long.In order to solve the problem of the long execution time of the framework,we propose to optimize the data transmission part of the framework by using matrix decomposition and the encoding technique based on the length of the attribute run,so as to shorten the running time of the framework.Experiments show that the above two methods can effectively reduce the running time of the framework.