The Application of Artificial Intelligence in Data Preprocessing

This study is based on the application of a real-measured data preprocessing method using data augmentation techniques. In response to the scarcity of sample data in the fields of construction and transportation, this paper proposes a data augmentation method based on the conditional generative adversarial network (CGAN). This method can learn the latent distribution of original data through a neural network and generate virtual data based on this distribution and specific conditional information, thereby meeting the training requirements of complex predictive models. Experimental validation was conducted on two data augmentation methods, introducing data augmentation tools from image recognition into the rectangular data in the construction and transportation fields for experimental analysis. The research results indicate that this data augmentation method can improve the accuracy and reliability of predictive models.