Unsupervised Learning
One thing I had not attempted yet in Machine Learning was unsupervised learning. I decided to use Fashion-MNIST again (usuing only 2 classes which were pullovers and jackets) and try some basic techniques on this data set with scikit-learn and PyTorch. There is a GitHub at the bottom linking to a notebook that is more…
Exploring PyTorch
The next project was completed using the Fashion-MNIST dataset and PyTorch to build a neural network. There is a GitHub link to a notebook with a more detailed version of this project provided at the bottom. PyTorch was something I had not previously explored so I decided this image classification problem would be a good…
An In Depth Project with Scikit-Learn
This next project I was tasked with taking a CSV all the way to a finalised model to predict with. The data provided was a regression problem with predicting insurance claim values. The data had a mixture of numerical values and categorical values. The data given was messy on purpose with lots of imputation needed…
Writing My Own Neural Network
The most technically complex problem I have worked on was designing my own Neural Network which I had to write without any packages. This was completed using only Numpy and Python. This project really allowed me to develop a deep understanding of Neural Nets and how they work and was a great learning curve for…
Predicting Film Meta Data from Scripts
The largest and most complicated problem I have worked on was my dissertation project which was a supervised learning classifier that used scripts to try and classify meta data such as genre, box office, age rating and awards success. The project involved web scraping data, saving the data to a SQL database, pre-processing this data,…
Follow My Blog
Get new content delivered directly to your inbox.
Nathan Brown Portfolio 