Creation and verification of an interpretable machine learning model for forecasting in-hospital results in Cardiorenal syndrome patients in critical condition