Model training method and device, additional feature data obtaining method and device, equipment and medium

Abstract

The invention provides a model training method and device, an additional feature data obtaining method and device, equipment and a medium. Obtaining a feature prediction model, the feature predictionmodel being trained based on the source data set, the feature prediction model being used for predicting at least one part of unique features of the source data based on at least one part of common features between the source data and the target data; obtaining a target data set; for each part of target data in the target data set, inputting at least one part of common features in the target datainto the feature prediction model to obtain at least one part of unique features predicted by the feature prediction model for at least one part of input common features; and taking at least one partof the predicted unique features as additional feature data of the target data. According to the method, the common features are taken as springboards, the unique features of the source data are migrated to the target data in a model migration mode, and the source data and the target data are not contacted in the process, so that the leakage risk of the source data is greatly reduced.

Description

technical field [0001] This application claims the priority of the Chinese patent application with the application number 201810929755.8, the application date is August 15, 2018, and the title is "Method, device, equipment and medium for model training and obtaining additional characteristic data". The present invention generally relates to the field of data science, and more specifically, relates to a method, device, device and medium for model training and acquisition of additional feature data. Background technique [0002] The goal of transfer learning is to transfer the knowledge acquired from the source data to the target data, so as to improve the use effect of the target data. [0003] Existing transfer learning algorithms are usually based on the circulation of data, and the source data needs to be brought into the environment of the target data during the implementation process. Since the migration process will touch the source data, the existing migration process...

AI Technical Summary

Problems solved by technology

Since the migration process will touch the source data, the existing migration process may cause the privacy of the source data to be leaked, making it difficult to guarantee the data security of the source data

Therefore, many data owners, such as banks, insurance, medical, financial and government departments, are unwilling to open their own data as source data, which greatly hinders the application and development of transfer learning technology

Images