View of Self-Supervised Credit Scoring with Masked Autoencoders: Addressing Data Gaps and Noise Robustly | Journal of Computer Technology and Software