Machine Learning Projects

Supervised + Unsupervised learning using SciKitLearn.

Project 1

Digit recognition using the MNIST dataset and SciKitLearn

1. Applying basic SKLearn libraries (Naive Bayes (incl. Gaussian Naive Bayes), Linear Regression, Logistic Regression, K-Nearest Neighbors) to identify object classes.
2. Functions to print grids of images.
3. Creating from scratch a bespoke image blurring function.

Project 2

Basic natural language processing using altnet forum messages and SciKitLearn

1. Applying basic SKLearn libraries (Naive Bayes, Linear Regression, Logistic Regression) to text data for message category prediction.
2. Basic text pre-processing using NLTK and Regex.
3. Creation of vocabularies for training machine learning models to predict message categories.

Project 3

Unsupervised learning using SKLearn and the Kaggle Poisonous Mushrooms dataset

1. Applying dimensionality reduction (principal component analysis) and basic unsupervised learning (Gaussian mixture models) to classify a highly dimensional dataset containing mushroom information to determine whether mushrooms are poisonous or non-poisonous.
2. Optimizing Gaussian mixture models.