Prediction method of unbalanced data set based on generative adversarial network

Abstract

The invention discloses an unbalanced data set prediction method based on a generative adversarial network. The method comprises the following steps: receiving a prediction request; collecting data to form a data set, and determining features and labels in the data set and the number of minority class samples and majority class samples; converting a non-numerical feature column and a label column in the data set into classification numerical values; standardizing the processed data set, and separating majority class samples and minority class samples in the standardized data set; synthesizing minority class samples a by using a deep convolutional adversarial network to form a balanced data set; dividing the balanced data set into a training set and a test set; constructing a convolutional neural network, and training the convolutional neural network by using the divided training set to obtain a trained convolutional neural network; finally, inputting the test set into the trained convolutional neural network to obtain a prediction result. According to the prediction method, the problem that minority class samples are very difficult to generate or even cannot be generated when big data are processed in an existing method is solved.

Description

technical field [0001] The invention belongs to the technical field of prediction methods for unbalanced data sets, and relates to a prediction method for unbalanced data sets based on generative confrontation networks. Background technique [0002] With the rapid development of information technology, data in various fields are being generated at an unprecedented speed and widely collected and stored. How to realize the intelligent processing of data and use the valuable information contained in the data has become a research hotspot in theory and application. Machine learning is a mainstream intelligent data processing technology. The classification problem is one of the important research contents in the field of machine learning. Some existing classification methods are relatively mature, and they can generally achieve better results in classifying balanced data. However, the data in the real world often have unbalanced characteristics, that is, the number of samples of ...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide a prediction method based on unbalanced data sets based on generative confrontation networks, which solves the problem that existing methods are very difficult or even impossible to generate minority samples when processing large data

Images