Integrated pruning strategy based on balanced binary tree

Abstract

The invention provides an integrated pruning strategy based on a balanced binary tree, and the structure of a multi-step multi-line ship lock comprises the following steps: S1, initializing a base classifier integrated pool: segmenting a big data set to form a plurality of sub-data sets, and carrying out the training and testing of each sub-data set, thereby forming an initial complete classifierpool; S2, constructing a balanced binary tree to form a final sub-integration; and S3, predicting and classifying new data samples by using the reserved optimal sub-ensemble. According to the method,the technical problems that a fitting phenomenon is easy to generate, a base classifier with too high or too low test precision is difficult to remove, and the generalization performance is not high are solved.

Description

technical field [0001] The invention relates to the technical field of integrated learning, in particular to an integrated pruning strategy based on a balanced binary tree. Background technique [0002] Integrated learning solves many problems faced by single classifiers in the process of massive data training and learning. However, since integrated learning consists of multiple single classifiers to form an integrated pool to complete prediction or classification tasks, it poses a higher requirement for computer hardware resources. Requirements, a common way to solve this problem is to use the idea of ​​integrated pruning strategy to reduce the number of single classifiers used as much as possible while ensuring that the final prediction or classification accuracy of integrated learning is not reduced or even improved. [0003] The current popular ensemble pruning strategy is the ensemble pruning strategy based on the clustering algorithm. This method uses the test accuracy...

AI Technical Summary

Problems solved by technology

[0005] The present invention provides an integrated pruning strategy based on a balanced binary tree, the purpose of which is to solve the problems in the background technology that the fitting phenomenon is easy to occur, it is difficult to eliminate the base classifier with too high or too low test accuracy, and the generalization performance is not high

Images