Gradient improvement decision neural network classification prediction method

Abstract

The invention discloses a gradient improvement decision neural network classification prediction method, and belongs to the technical field of big data analysis and machine learning classification prediction. The gradient improvement decision neural network classification prediction method comprises the following steps: firstly, establishing a data sample set according to the characteristics of apredicted object, setting 70%-90% of samples in the data sample set as a training set, and taking 10%-30% of samples as a training set; and secondly, proposing a gradient improvement decision neural network classification prediction method for training and testing the data sample set. The gradient improvement decision neural network classification prediction method has the advantages that throughthe feature selection and feature expansion functions of the gradient lifting tree, the attribute features of the samples are increased, and the problem that the prediction precision is low when few feature data samples are processed through an existing algorithm is solved.

Description

technical field [0001] The invention belongs to the technical field of big data analysis and machine learning classification prediction, in particular, it provides a gradient boosting decision-making neural network classification prediction method, which is suitable for solving various classification and regression problems, and can be applied to the fields of data analysis, evaluation and fault prediction . Background technique [0002] In recent years, artificial intelligence has achieved unprecedented development, and various intelligent algorithms such as machine learning, deep learning, and neural networks have been widely studied and applied, and achieved good results. However, with the increase of data volume and the continuous improvement of users' requirements for data prediction performance, the existing prediction algorithms can no longer meet the above requirements, and there is an urgent need to improve existing algorithms and propose new prediction algorithms. ...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a gradient boosting decision-making neural network classification prediction method to solve the problem of low accuracy of existing prediction algorithms when processing classification predictions with few feature samples

Images