Your stupid machine doesn't even know how to read. So we have to teach it how to do it.
More spesifically, we're going to learn it how to use Python to recognize handwritten digits, with a little help from the MNIST dataset, Scikit Learn and the K Nearest Neighbor algorithm.
For this tutorial, you'll need the following installed on your computer:
- Numpy
- Pandas
- Scikit Learn
- Scipy
- Matplotlib
If you already have pip installed, simply type the following into your command line interface:
sudo pip install numpy pandas scikit-learn scipy matplotlib
Video tutorial
In the first part of the tutorial, we’ll have a look at the dataset, how it’s structured and how the computer can interpret it. In addition to this, we’ll also have a look at our algorithm of choice, which is called K Nearest Neighbor:
In the second part, we’ll get your dependencies up and running, import the dataset and get going with the coding. We’ll train the algorithm with a few different sizes of data and have a look at how its success rate improves: