Feature importance sorting system based on random forest algorithm in multi-center mode

摘要

Disclosed is a multi-center mode random forest algorithm-based feature importance sorting system, comprising a front-end processor deployed in each center participating in collaborative computing, a central server receiving and integrating feature importance sorting results of the various centers, and a result display module feeding back a final feature importance sorting result to a user. A feature importance sorting result is respectively calculated at each center according to a multi-center-based random forest algorithm, and the central server integrates the sorting results of the various centers to form a global feature importance sorting result. The present invention operates under the condition that data in the various centers is not exposed, such that the data in the various centers remains in the various centers, only intermediate parameters are transmitted to the central server, and the original data is not transmitted, so as to effectively ensure data security and the personal privacy included in the data.