The repository was created to show some initial results of my original work on the new classification algorithm. The code must remain private as the algorithm has not been published yet.
Files:
- MSc-Results.pdf - Contains the comparison chapter of my master thesis in which the algorithm is compared with Naïve Bayes, Decision trees, Multi-Layer Perceptron, and SVM. The results are calculated as a macro average of 5-fold validation repeated 10 times.
- showcase.ipynb - Shows exemplary training results and classification quality obtained for validation sets for multiple datasets, including two generated datasets and four popular benchmark datasets.