Image Classification on MNIST original unenhanced (test)
98.6Accuracy (%)2-layer NN
Evaluation Results
| Method | Links | |
|---|---|---|
| 2-layer NNhidden_units=800, loss_function=Cross-Entropy Loss2018.04 | 98.6 | |
| Tsetlin Machinestatistic=95 %ile2018.04 | 98.3 | |
| Tsetlin Machinestatistic=5 %ile2018.04 | 98.1 | |
| K-nearest-neighborsdistance_metric=L32018.04 | 97.2 | |
| 3-layer NNhidden_units=500+1502018.04 | 97.1 | |
| PCA + quadratic classifierpca_components=402018.04 | 96.7 | |
| RBF + linear classifierrbf_units=10002018.04 | 96.4 | |
| Logistic regression2018.04 | 91.5 | |
| Linear classifierarchitecture=1-layer NN2018.04 | 88 | |
| Decision tree2018.04 | 87.8 | |
| Multinomial Naive Bayes2018.04 | 83.2 |