Missing value completion method based on generative adversarial network and medium

Abstract

The invention discloses a missing value completion method based on a generative adversarial network and a medium. The method comprises the steps of: collecting incomplete historical data, and obtaining a training set; iteratively training a generator and a discriminator in an offline mode based on the training set; determining an adversarial network model through a cross validation method, and obtaining an optimal generator with the minimum error; and based on the optimal generator, carrying out missing value inference on the acquired data, and reconstructing complete data. The problem that a data driving method needs a large amount of complete historical data is solved, and the requirement of a traditional missing value inference method for a data structure is also avoided. The method is mainly used for solving the problem of data quality reduction caused by loss of collected data due to faults of collection equipment, insufficient coverage of the collection equipment, errors of operators and the like.

Description

technical field [0001] The invention belongs to the field of data cleaning, and in particular relates to a missing value complement method and medium based on a generative confrontation network. Background technique [0002] As an important production factor, data is the basis of scientific research analysis and many applications, and its quality is directly related to the effect of the model and the final result. Only complete and accurate data can achieve ideal results, while missing and abnormal data can even lead to wrong conclusions. However, many reasons will lead to data loss, such as failure of collection equipment, insufficient coverage of collection equipment, operator error, etc., so data loss is inevitable. Therefore, in order to obtain correct results, data cleaning is an indispensable part of the entire data analysis process, and missing value completion is an important part of data cleaning. [0003] At present, the commonly used missing value completion met...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide a missing value completion method and medium based on generative adversarial networks, which are used to solve the loss of collected data caused by collection equipment failure, insufficient coverage of collection equipment, operator errors, etc. The problem of quality reduction, does not require complete training data and can more accurately fill in missing values

Images