- Abubakr Bhula
- Aum Patel
- Muhammad Azhar Gaffar
League of Legends Diamond Ranked Games (10 min)
This is the final repository for the IN3062 Coursework. We are using this Dataset to categorise whether a game of League of Legends is a Win or Loss for the Blue Team based on the first 10 minutes of a game.
Task | Description | Done by |
---|---|---|
Analysis of Dataset | We have analysed the Dataset in order to look for null values, consider correlation and look at other factors that we may need to take into account. | Everyone |
Logistic Regression | Implementation of Logistic Regression with improvements. | Abubakr |
Decision Trees | Implementation of Decision Tree Classification algorithm with improvements. | Aum |
K Nearest Neighbors | Implementation of KNN Classification with improvements. | Muhammad |
Neural Network | Implementation of a Multilayer Perception Model with improvements. | Everyone |
- The accuracy notes at the bottom of each model was used to track our workflow for each model in preparation for the report. You will find the final accuracy scores in the report- they have been standardised with each model by using the same "Random State" split as a control.
- In KNN model, line 128 and 135 have been commented out as they take a long time to run. I would recommend uncommenting them if you would like to see our baseline model and a plot of the n_neighbors tested. This would be the full output we show in the report.
- In the Neural Network model, line 129 and 132 have been commented out as they take a long time to run. I would recommend uncommenting them if you would like to see our baseline model and intermediate model.