Lab 8

We will create dynamic endpoints from the SVM endpoints. Remember this is focused around using the built in IRIS dataset. If you need a refresher please look back at the lecture from 4-16-2020. IRIS Lecture 4-16-2020

1. How does changing the splitting change the accuracy?.

a) Create dynamic endpoints that change the splitting within the function train_test_split b) make a table in markdown format that has the results (accuracy) for 60,70,75,80,85,90 % training split

2. Create a dynamic endpoint of your choice

Provide the enpoint and a description here:

3. Discuss pros, cons and ways forward to this type of implementation

i.e. machine learning flask app, in a container. Think about scaling.

4. Put this in a container

I have done this for you so now answer these questions:

a) What do I need to change in the Dockerfile if I want to enter the shell interactively with the make docker-alli command?

b) What are the benefits of running this in a container ?

c) Explain the make docker-all commands in the makefile