Skip to main content
. 2022 Jan 25;14(1):2026208. doi: 10.1080/19420862.2022.2026208

Figure 2.

A 2 × 2 plot showing the correlation coefficients of experimental aggregation rates and predicted aggregation rates of the linear model (top) and the KNN model (bottom). On the left is the correlation using all 20 data, and on the right is the correlation using LOOCV.

Correlation coefficients for the best two-feature linear, support vector regression (SVR) and k-nearest neighbors (KNN) regression models trained using all 20 data and LOOCV. The features for the linear regression model are SCM_neg_H2 and SASA_phobic_H3. The features for the SVR and KNN models are both SCM_pos_H2 and SASA_phobic_Fv.