Electronic Medical Records (EMRs) contain a large number of missing values which imposes difficulties for data scientists who want to model after this data. In a previous post, we discussed the different feature engineering methods available on diagnosis codes, medication data and clinical notes of EMRs. In this post, we