An open API service indexing awesome lists of open source software.

https://github.com/solegalli/udemy-feml-challenge


https://github.com/solegalli/udemy-feml-challenge

Last synced: 5 months ago
JSON representation

Awesome Lists containing this project

README

        

## Feature Engineering for Machine Learning - Course Assignment

![Python 3.7](https://img.shields.io/badge/Python-3.7-blue.svg)

Published November 2019

[](https://www.udemy.com/course/feature-engineering-for-machine-learning/?referralCode=A855148E05283015CF06)

## Links

- [Course Page](https://www.udemy.com/course/feature-engineering-for-machine-learning/?referralCode=A855148E05283015CF06)

Final challenge / assignment for the Udemy course: Feature Engineering for Machine Learning

## Assignment:

Beat the performance of my Lasso regression by using different feature engineering steps **ONLY!!**.

The performance of my current model, as shown in this notebook is:

test mse: 1063016789.3316755
test rmse: 32603.938248801718
test r2: 0.8453144708738004

To beat my model you will need a test r2 bigger than 0.85 and a rmse smaller than 32603.

================================================================================================

### Conditions:

- You MUST NOT change the hyperparameters of the Lasso.
- You MUST use the same seeds in Lasso and train_test_split as I show in this notebook (random_state)
- You MUST use all the features of the dataset (except Id) - you MUST NOT select features

================================================================================================

### If you beat my model:

Make a pull request with your notebook to this github repo:
https://github.com/solegalli/udemy-feml-challenge

And add your notebook to the folder:

-StudentsSolutions_v1.0.0

### How to make the PR

1) fork the repo:

Go to https://github.com/solegalli/udemy-feml-challenge, and click on the **fork** button at the top-right

2) clone your forked repo into your local computer:

- Go to www.github.com/yourusername/udemy-feml-challenge
- Click the green button that says clone or download
- copy the url that opens up
- power up a git console
- type: git clone (paste the url you copied from github)
- done

3) Make a copy of the jupyter notebook and add your name:

- Open up the Jupyter notebook called 13-Assignement.ipynb
- Click the "File" button at the top-right and then click "Make a copy"
- **Work your solution in the Copy** and not in the original assignment (otherwise there will be conflicts when making the PR)
- Change the name of the copy of the notebook to: 13-Assignement_yourname.ipynb
- Move the notebook to the folder **StudentsSolutions_v1.0.0**
- done

When you finish, just commit the new notebook to your fork and then make a PR to my repo.

- git add StudentsSolutions_v1.0.0/13-Assignement_yourname.ipynb
- git commit -m "your commit message"
- git push origin master or git push origin yourfeaturebranch
- go to your repo and make a pull request.

## But i have a notebook ready and I haven't cloned the repo yet, how can I make the PR?

If you worked in the copy you downloaded from Udemy before forking and cloning this repo, then follow this steps:

1) fork the repo:

Go to https://github.com/solegalli/udemy-feml-challenge, and click on the fork button at the top-right

2) clone your forked repo into your local computer:

Go to www.github.com/yourusername/udemy-feml-challenge

- Click the green button that says clone or download
- Copy the url that opens up
- Power up a git console
- Type: git clone (paste the url you copied from github)
- Done

3) Rename your solution as follows and copy it into your cloned repo:

- Rename your solution notebook to: 13-Assignement_yourname.ipynb
- Copy this file into the cloned repo, inside the folder **StudentsSolutions_v1.0.0**
- Done

When you finish, just commit the new notebook to your fork and then make a PR to my repo

- git add StudentsSolutions_v1.0.0/13-Assignement_yourname.ipynb
- git commit -m "your commit message"
- git push origin master or git push origin yourfeaturebranch
- go to your repo and make a pull request.

**Good luck!!**