Title：Recognition system, device and method based on gradient boosting decision tree
Country：China
Patent No.：201911216442.9
Legal Status：Authorized
Inventor：Wei Song, Milin Zhang
Assignee：Tsinghua University
Address：Tsinghua University,Haidian District Beijing 100084, China
Filing Date：2019-12-02
Issue Date：2022-12-2
Abstract：
The invention discloses an identification system, device and method based on a gradient boosting decision tree, and the method comprises the steps that a processing module carries out the parallel comparison of received signal features through an assembly line according to model parameters in a loaded gradient boosting decision tree model, obtains a plurality of total prediction scores, and transmits the total prediction scores to a result generation module; the result generation module generates an identification result according to the plurality of received total prediction scores. In the processing module, an assembly line is used for carrying out parallel comparison on received signal features, the processing speed can be increased, and hardware power consumption and processing time expenditure are reduced. The signal features are compared through the gradient boosting decision tree model, the recognition result is generated, mixed signal data can be flexibly processed, the robustness for abnormal points is high, and the recognition result with high accuracy can be obtained. Processing is divided into four stages, the controller is used for controlling operation of all the stages in the assembly line, parallel processing can be achieved in the FPGA, the processing speed is increased, and the universality is high.
Patent Certificate：
PDF/Jpg