ml-zoomcamp-exercises

Exercises performed as part of the ML Zoomcamp course. This course is conducted by Alexey Grigorev to run for 4 months starting from 6-Sep-2021.

• https://datatalks.club/courses/2021-winter-ml-zoomcamp.html
• https://github.com/alexeygrigorev/mlbookcamp-code/tree/master/course-zoomcamp

This repo has the exercises I did and any notes I took as part of this course. I have also prepared some guides for deployment of python web app to PythonAnywhere, Heroku, docker containers using Google Cloud shell.

Guides

• How to use PythonAnywhere to deploy Python Web apps
• How to use Heroku
• How to use Google Cloud shell for docker
• How to install Docker on Ubuntu/CentOS/Amazon Linux

Session #1

• Learning Numpy
• Learning Pandas

Session #2

• Car Price prediction using Linear Regression

Session #3

• Churn prediction using Logistic Regression for Classification
• Explore more - Accuracy using all features Vs removing least important features
• Explore more - Compare LinearRegression and RidgeRegression for car price data. Test different regularization values for RidgeRegression

Session #4

• Evaluation metrics for Classification
• Explore more - Precision and Recall for dummy model, Precision, Recall for different thresholds and more

Session #5

• Use PythonAnywhere to deploy your web service with ML model
• Use Heroku to deploy your web service with ML model

Session #6

• Trees and tree-based ensemble algorithms