Machine Learning program to predict soccer game results using the football-data.co.uk dataset for training.
Below shows the final error rates for the different learning methods used after optimization:
SVM Linear | SVM RBF | Polynomial Regression | Random Forests | SVM Poly | QDA | SVM Sigmoid |
---|---|---|---|---|---|---|
42.49% | 43.14% | 44.21% | 44.31% | 46.49% | 48.97% | 58.7% |
Read full setup, results and analysis here.
This program is dependent on the python libraries: xlrd for parsing the excel data files, scikit-learn and SciPy for implementing the machine learning algorithms.
To install python, download and install a python version from here.
Next, in order to download the required libraries, install pip following the directions at pypa.io.
Once pip is installed correctly, in order to install the xlrd library simply run the following command:
pip install xlrd
Next, install scikit-learn using the following command:
pip install sklearn
Lastly, install SciPy using the following command:
pip install scipy
You can download this project by using the following command:
git clone https://github.com/kvelcich/Soccer_Fixture_Predictor
To run the program, in the main directory of the project, run the following command:
python main.py
You can also dig around, editing the hyperparameters and the range of the inputs given to the algorithms in order to achieve differing results.