A Channel Environment Missing Data Reconstruction Method Based on Conditional Generative Adversarial Networks
Missing data reconstruction is an important part of channel environment data preprocessing,and its reconstruction effect directly affects the quality of channel environment monitoring and maintenance.However,the current methods for reconstruction are not ideal.In practical applications,the proportion of data reconstruction errors is relatively small,and the data reconstruction rate is relatively low.Therefore,a channel environment missing data reconstruction method based on conditional generative adversarial networks is proposed.Using descriptive statistics to identify missing data in the channel environment dataset,establishing a conditional generative adversarial network,training the channel environment data using the adversarial network,extracting missing data features,selecting the generated data that is closest to the actual situation to reconstruct the missing data,and achieving channel environment missing data reconstruction based on the conditional generative adversarial network.Experimental results have shown that the design method has effectively improved the proportion of data reconstruction errors to zero,and the reconstruction rate has also been effectively improved.It has good application prospects in the reconstruction of missing data in channel environments.
NETL
NSTL
万方数据
