Table 2 Machine learning methods performance evaluation.
From: Improved diagnosis of rheumatoid arthritis using an artificial neural network
Cross-validation accuracy (± SD) | Cross-validation AUC (± SD) | |
|---|---|---|
ANN (1 hidden layer) | 0.901 ± 0.014 | 0.945 ± 0.018 |
ANN (2 hidden layers) | 0.907 ± 0.022 | 0.948 ± 0.016 |
Logistic Regression | 0.903 ± 0.013 | 0.947 ± 0.015 |
Random Forest | 0.897 ± 0.019 | 0.937 ± 0.010 |
K nearest neighbors | 0.879 ± 0.013 | 0.924 ± 0.012 |
Support vector machine | 0.901 ± 0.014 | 0.890 ± 0.015 |
Gaussian Naïve Bayes | 0.872 ± 0.020 | 0.942 ± 0.013 |
Gradient boosting classifier | 0.900 ± 0.027 | 0.948 ± 0.009 |
