Table 3 Performance measures for assessing the predictive accuracy of mania treatment continuation prediction models. Single-year data are from 2016, and two years of data are from 2013–2015
From: Unlocking treatment success: predicting atypical antipsychotic continuation in youth with mania
Model | AUC | Intercept | Slope | |||
|---|---|---|---|---|---|---|
(single year) | (consecutive 2-year) | (single year) | (consecutive 2-year) | (single year) | (consecutive 2-year) | |
Generalized Linear Model (Reference) | 0.82 (0.79–0.85) | 0.82 (0.79–0.85) | -1.24 | -1.24 | 1.00 | 1.01 |
Support Vector Machines with Linear Kernel | 0.82 (0.79–0.85) | 0.82 (0.79–0.85) | -1.23 | -1.24 | 1.14 | 1.14 |
Generalized Additive Model using LOESS | 0.82 (0·78–0·86) | 0.81 (0·78–0·85) | -1.25 | -1.25 | 1.12 | 1.12 |
Stochastic Gradient Boosting | 0.81 (0·75–0·81) | 0.81 (0·75–0·81) | -1.26 | -1.26 | 1.19 | 1.17 |
Random Forest | 0.78 (0·75–0·81) | 0.77 (0·75–0·80) | -0.29 | -0.28 | 1.78 | 1.76 |
Super Learner | 0.82 (0·79–0·85) | 0.82 (0·79–0·85) | -1.24 | -1.24 | 1.01 | 1.00 |