Prediction model of Hepatorenal Syndrome
Hepatic renal syndrome (HRS) is a type of kidney failure that occurs in progressive cirrhosis (liver dysfunction). The three-month survival rate of HRS patients is 15%. However, currently, biomarkers , Sedimentation factors, and useful prognostic criteria have not yet been discovered. There is also no correlation with liver function such as Child-Pugh classification. Fortunately early treatment is effective, and early prediction of HRS is important. In this study, using random forest, HRS onset in the medical record of cirrhosis patients at admission from the MIMIC 2 public database Predicted. As a result, our prediction model achieved AUC of 0.81. This prediction is the first prediction of HRS and it has become> so that it can be done from an extremely early stage. This research was papers and Nakano of our company announced orally at the workshop of NIPS 2015 which is one of the world's best machine learning international conferences.