This project contains the code for this video on Youtube by Siraj Raval and is aimed to teach the concept of "Gradient Descent". We are using a small dataset of student test scores and the amount of hours they studied. Intuitively, there must be a relationship right? The more you study, the better your test scores should be. We're going to use linear regression to prove this relationship.
The target is to predict the students test score, when given the amount they studied 📖🤓📚.
I've generated a preview file, which allows us to view the jupyter notebook without any further installation steps.🙌
(Hint: I've done this by running jupyter nbconvert --to markdown Linear_regression.ipynb)
- Clone this repository
- Install the dependencies
- Open the
Linear_regression.ipynbfile from the browsers jupyter tap.
We'll need Python 2 or 3 with the following dependencies:
- numpy ... (for matrix multiplications)
- matplotlib ... (for creating visualizations)
- jupyter ... (for starting the jupyter notebook)
I've used conda to create an virtual environment and to install the dependencies.
# create a virtual environment with required dependencies
conda create --name tutorial-env numpy matplotlib jupyter
# activate the environment
source activate tutorial-env # for macOS and Linux Users
# run the jupyter notebook server
jupyter notebook
# ... edit the notebook in your browser
# When done, shut down jupyter notebook and deactivate the virtual environment
source deactivate tutorial-env # for macOS and Linux User
Of course you can also use pip to install any dependencies.
Credits for this code go to mattnedrich and Siraj Raval. I've merely created a summary for myself.