Query: AI for Managers


#1

[ From internal discussion]
Had a few queries regarding the AI for Managers course.

  1. I practised the housing data exercise in Azure ML. I was confused, which dots to join? I had to look at the original video for everything. Would like to understand the logic behind it. Eg why is normalize data first dot joining with train model & cross validate but not with linear regression?

  1. How do I validate the results from the above housing data study?

  2. Any other test sets I can test this on?

  3. I was one of the early enrollers for the course. Unfortunately, last few months were really busy at my end & I couldn’t complete the course. I’m now switching jobs & have 7-10 free where I would like to focus on the course. My Boot ML’s subscription is over & I havent been to test the execrise in Boot ML. Can you help extend my subscription for the lab?

Look forward to hearing from you.


#2

The idea is simple. Split data into two parts: train and test. Train the model on train set and then test its performance on test. The basic cleanup of data such as missing data and normalization is done in both cases. Once we have trained a model, we further improve its performance by fine tuning the hyper parameters

Basically, we usually train the mode using cross-validation.

You can do the predict on your dataset.

Yes. You can drag and drop your dataset.

okay.