recommendation system for eCommerce healthcare portal suggestion

Question

I am trying to build a recommendation system. My system is basically a ecommerce application where our customers answers a bunch of questions related to healthcare (their basic health related question). Based on their ansers, we recommend some product. This process of recommendationis based of conventional rule base aapproach. Think of it as bunch of if-else condition. Now I am playing around with some machine learning technique nd want to see if this way will add any value in our healthcare system. I am at the very starting point and can use any suggestion from you guys. The suggestion could aim towards following:

1. Any product that you feel which leverages ML technique with respect to health(considering HIPPA constraints)
2. Any product that you feel which leverages ML technique with respect to health
3. What could be the first step towards building such a system.

Answer 1

Your recommendation system will be designed to tell the customer what product they should choose, however, this doesn't account for what products the customer likes. A ML method could take all the input parameters from the recommendation system and provide a recommended product based on what similar users liked.

There's no specific ML technique for considering HIPPA constraints. This would come into play more in the pre-processing stage. For example, you might not be able to acquire specific birth dates and addresses, but maybe you can get an age range and a zip code (first 3 digits only).

Logistic regression is frequently used in healthcare for binary classification.

There are lots of tutorials online for building ML models. Kaggle has a good tutorial for Python and R using titanic survival data.