Training method and device based on machine learning model

Abstract

The invention discloses a training method and device of a machine learning model. The method comprises the steps of performing single training on the current machine learning model based on each training sample; obtaining an initial feature contribution degree set of each training sample of which the training result meets a preset condition, determining a target feature contribution degree in eachinitial feature contribution degree set, wherein the target feature contribution degree in the target feature contribution degree set is greater than other feature contribution degrees in the initialfeature contribution degree set; based on each target feature contribution degree set, constructing a target training sample; and training the current machine learning model based on the target training sample. According to the training method, the target features with the large feature contribution degree are reserved in the target training sample; the influence on the training accuracy of the current machine learning model is small; and compared with manual screening, the speed is high, and the training accuracy of the model is not affected.

Description

technical field [0001] The invention relates to the technical field of data processing, in particular to a training method and device based on a machine learning model. Background technique [0002] In the training process of the current machine learning model, although the redundant features in the training samples will not affect the performance of the model, it will make the computer do useless work. In the existing technology, in order to reduce the computational pressure in the training process of the machine learning model, based on thousands of features in the training process, the screening of features usually relies on manual screening. [0003] Manual screening not only takes a long time, but also the accuracy of the screening cannot be guaranteed, resulting in low accuracy of machine learning model training when the machine learning model is trained based on the filtered data. Contents of the invention [0004] In view of this, the present invention provides a ...

AI Technical Summary

Problems solved by technology

[0004] In view of this, the present invention provides a training method and device based on a machine learning model to solve the problem of manual screening of features during the training of the machine learning model, which not only takes a long time, but also has low training accuracy.

Images