Fig. 1From: A comparative study: classification vs. user-based collaborative filtering for clinical predictionThe simulation pipeline proceeds in a series of steps: (1) The data is input, and (2) divided into K-folds, (3) missing data is created at random, (4) the data is discretized, (5) a recommender system is fit along with variations of logistic regression and random forests, and the (6) performance is evaluatedBack to article page