Machine Learning in Healthcare is a rich field, but it is challenging. Much of the data is highly protected. Generally, the data is poorly organized and requires significant effort to transform into anything more than Garbage In, Garbage Out.
Start with a Data Warehouse
How to get the clean data? You need to build a quality Data Warehouse - Old School Kimball Style.
When the beach is clean of garbage, we can see the patterns in the sand, the grasses, the stones, and the waves.
Photo by Tobias Tullius on Unsplash
Learn the Hell out of an ML Algorithm
Machine Learning is not Mathemagic - its applied statistics. Pick a model and learn the hell out it. It doesn't matter if it applies to customers, patients, flowers in a field... Math is math. You can't make magic out of math, and math doesn't exist in a vacuum, but in order to solve the business problem, you need to know the math.
The CLV Algorithm Medium Articles
A very important model that has general healthcare implications is the CLV model.
Learn by doing. Writing this, I don't know what comes next, but I will have faith that by doing it, I will understand and see where to go. Learning is a journey of faith.
History of CLV
According to Wikipedia, the CLV is a model to predict the lifetime value of a customer.
References to the model first appeared in Shaw and Stone, Database Marketing.