Page 36 - Kỷ yếu hội thảo khoa học lần thứ 12 - Công nghệ thông tin và Ứng dụng trong các lĩnh vực (CITA 2023)
P. 36

20




















































                                     Fig. 1. The RMSE performance with various missing rates



                     5     Conclusion

                     In this paper, we present a different method to handle the missing data named Evolu-
                     tionary Generative Adversarial for Imputation Data (EGAIN). This novel framework
                     takes advantage of GAN frameworks and evolutionary computation paradigm, which
                     allows more than individual adversarial objectives, and selects the best candidate for
                     each training iteration. Experiments highlights the good performance of EGAIN against
                     other methods.


                     References


                      1.  M. Arjovsky and L. Bottou. Towards principled methods for training generativeadversarial
                         networks. arXiv preprint arXiv:1701.04862, 2017.




                     CITA 2023                                                   ISBN: 978-604-80-8083-9
   31   32   33   34   35   36   37   38   39   40   41