Future improvements

Our project certainly has a few possible future improvements that are worthy of being implemented. These include:

• Using a voting regressor, which uses several distinct models to 'vote' for property prices, allowing us to amalgamate several models together.
• Developing additional methods for model explainability, such as visualisations of the data points for kNN regressor.
• Ensembling other base regressors besides decision trees.
• Adding hyperparameter tuning for our rolling models, which may improve their performance metrics.