Ok so far we have seen how to make Machine Learning models, we explored many datasets and also gained valuable insights from these datasets. This is valuable and it showed how Exploratory Data Analysis can be performed. But we are not done yet, until we use these models in any application. Let’s see this with examples:

1. Suppose you are building ML based recommender system. You ask end user their preferences and based on those you recommend them the movie etc. While basic analysis can be done on Jupyter Notebook…but this fails when you must predict values based on inputs from the end user.
2. Suppose you are a doctor. You want to know if the person has chronic Hepatitis or not and based on the medical condition of the patient you want to know if this patient is going to live or not. Now this is classical Classification problem. As a doctor you have limited time, and notebooks are not feasible option for you. Hence having an interface to interact with this model is vital.
3. Suppose you are a business owner. You want to see how will your salespeople perform in the next quarter. You will need to have some application to predict these values.

From above examples we can see we need to expose these models effectively in order to make them useful. In this post lets explore how to share these models in applications.

#data-science #joblib #python #machine-learning #machine-learning-models