With the mortality rate increasing to 71.3% once adults reach 40, Cirrhosis is a disease of concern for 85.6% of adults who drink. Thus, we made ML models to predict the efficacy of D-penicillamine.
Alcohol abuse is a prevalent issue in the United States as nearly 15 million people above 12 suffer from it. It is especially an issue for young adults in college who often go binge drinking. Because of this, many liver issues can arise, and Cirrhosis is a dangerous one. Thus, we wanted to predict whether drug D-penicillamine would be successful in increasing the chances of survival.
We built machine learning models to predict the efficacy of D-penicillamine based on various features including Age, Sex, Patient Health Information, and Stage.
We cleaned up the data, and used the decision tree model and K Nearest Neighbors.
We had limited time to code because we switched our project very late. We tried to remove the outliers and scale the features, but that removed too much data and made our model less accurate.
We were able to build a model with over 80% accuracy rate with limited time.
We learned what the decision tree model is, as well as ways to improve it.
We can use our existing code and test it out on different drugs to find the best drug to treat Cerrhosis.
Log in or sign up for Devpost to join the conversation.