Distributed random forest evaluation system and method with privacy protection attribute

Abstract

The invention discloses a distributed random forest evaluation system and method with a privacy protection attribute. The system comprises a user and a random forest evaluation platform, wherein the evaluation platform is composed of t evaluation servers, and a decentralized distributed structure is adopted. The data evaluation method comprises the following four parts: initializing a system; sending user information; evaluating data in a random forest manner; and decrypting the evaluation result. According to the invention, random forest assessment is carried out on the encrypted user data under the condition that the user data and the decision tree model of the server side are not leaked, and only the target user can obtain the assessment result. Even if a small number of fault servers exist in the evaluation process, it can be ensured that the evaluation platform outputs a correct result. Therefore, the method and the system have very high robustness and practicability.

Description

technical field [0001] The invention belongs to the technical field of cryptography and information security, and relates to a distributed random forest evaluation system and method with privacy protection attributes. The system uses t servers to perform random forest evaluation on encrypted data of users, and only target users can obtain the evaluation As a result, the data privacy of data providers and model holders is protected. Background technique [0002] With the rapid development of computer technology, machine learning algorithms have been practiced in more and more fields such as object detection, image classification, and disease diagnosis, and have achieved remarkable results. The effect of a machine learning model not only depends on the pros and cons of the machine learning algorithm, but also needs to use massive data to train and test the model. But in practice, data is often scattered everywhere and difficult to centralize. Data holders may only perform mo...

AI Technical Summary

Problems solved by technology

[0006] 1) The solution requires the use of an outsourced server;

[0007] 2) The efficiency of the fully homomorphic encryption algorithm is relatively low;

[0008] 3) The robustness of the scheme is not strong. Whether it is an outsourced server or any model holder that fails, the system will not be able to output the evaluation and prediction results;

[0009] 4) The predicted result can only be decrypted by a specific user, and has no flexibility in a multi-user environment

Images