I am an econometrics student with a keen interest in combining traditional econometrics with mordern machine learning algorithms. This repository is where I will be posting some of my projects as I explore my interest further.
Projects:
- Predicting Housing Prices with Random Forests (my first!!)
- Modelling Prepayment Rates of Fixed Rate Mortgages - (using Linear Panel Model, Random Forests, SARIMAX)
- Sentiment Analysis of Movie Reviews
- Scam detection for Ads on Facebook marketplace
- Simple Trading Strategies using time series modelling - work in progress: until now I have come up with a way to detect pairs.
- Long Document Classifier with Simple Chunking Mechanism (warning: the model acheives 100% accuracy on test, only because the task was really easy and the model is very complex. My main aim was to only show how to fine-tune BERT with document sizes longer than its max 512 tokens. In the project I am currently working on in my internship, I also use BERT + Chunking but for a very complex task and I am not even close to 100% accuracy!)
Feel free to explore and reach out if you have any questions or suggestions!