Single-model fusion method based on cross validation

Abstract

The invention particularly relates to a single-model fusion method based on cross validation. The single-model fusion method based on cross validation adopts a basic machine learning method and comprises the steps of data preprocessing, data set division, cross validation, model training, prediction and model fusion. According to the cross validation-based single model fusion method, cross validation is utilized to improve model generalization, and a multi-model fusion idea is combined to mine data at different levels. Meanwhile, a model is relatively single. A result is generally relatively robust, and the prediction effect is superior to that of a single model or a common multi-model fusion method.

Description

technical field [0001] The invention relates to the technical field of model training and fusion in machine learning, in particular to a single-model fusion method based on cross-validation. Background technique [0002] In the wave of artificial intelligence development, the emergence of neural networks and deep learning has made a big step forward in the development of the industry, but the classic machine learning method benefits from its advantages of high training efficiency, less data required, and strong interpretability. It still plays an irreplaceable role in some fields. [0003] Currently, there are usually two types of classic machine learning models: single-model full-data training and multi-model fusion. The use of a single model usually uses all training sets for model training and prediction of the test set, while multi-model fusion is usually based on methods such as voting, averaging, bagging, boosting, and stacking of multiple homogeneous or heterogeneous...

AI Technical Summary

Problems solved by technology

[0008] 1. The generalization ability of a single model is not strong, and there may be incomplete observation of the data or overfitting

[0009] 2. The calculation of multiple models is complicated, and the difference in properties between different models is not easy to control, and it is not easy to obtain reliable results

[0010] 3. In the case of limited data, it is impossible to fully tap the potential value of the data

Images