6th March 2024
I have run the random forest and have my foundation ready. From manual analysis, we see that different oppositions need different time frames for the training and testing data, as well as different predictor variables. For example, England needs 2011, but France needs 2018.
Goals and background
I took data from Kaggle (https://www.kaggle.com/datasets/lylebegbie/international-rugby-union-results-from-18712022/), which gives me data for all the rugby union matches from 1871-2023.
In this project, I look to predict the results of the 2024 Guinness Six Nations tournament, in particular the score difference between a team and their opposition.
In addition, I also look to answer some questions:
- Is there such a thing as "home advantage"?
- How important is "momentum"?
I plan to use a Random Forest model as that makes the interpretation easy of the results, but I will also consider a LSTM model.