A repository for the Emerging Technologies assignments. These assignments will be made with Python and Jupyter Notebook.
This repository will consist of the following programs Python scripts and Jupyter notebooks.
A jupyter notebook explaining the concepts behind and the use of the numpy random package, including plots of the various distributions.
A jupyter notebook explaining the famous iris data set including the difficulty in writing an algorithm to separate the three classes of iris based on the variables in the dataset.
A jupyter notebook explaining how to read the MNIST dataset efficiently into memory in Python.
A Python script that takes an image file containing a handwritten digit and identifies the digit using a supervised learning algorithm and the MNIST dataset.
A jupyter notebook explaining how the above Python script works and discussing its performance.
Although Juptyer notebooks can be previewed on GitHub, I recommend you download these notebooks and run them locally as a lot of the code is reliant on generated local files and images.You can manually download the MNIST datasets and place them in the MNIST_data folder (create it if it doesn't exist) or you can use the download function of the script which will download and store the dataset for you.
The following files are a mix of Python files and Jupyter notebook files.
Python can be downloaded from here.
Once downloaded and installed, open a command prompt and run the following command to install Jupyter.
pip install jupyter
Once installed, download or clone this repository to your local machine.
Simply navigate to the folder your just downloaded/cloned and run the following command in the directory of the folder
jupyter notebook
This will run a Juptyer notebook on your local machine in your default web browser.
To open the Python script simply type
python digit-recognition.py
If you get an error regarding a unknow modules/errors with import please ensure you have the correct Python library installed. For example, Jupyter is not installed by default with Python but can be installed with the command found above. Similarly, if TensorFlow can't be found, you can simply run the same command but replace jupyter with TensorFlow. Good luck!
If you encounter any other problems please refer to this article.
The following is a collection of all the resources I used to develop and write these notebooks/scripts.