Skip to main content
Fig. 5 | Journal of Cheminformatics

Fig. 5

From: QSAR-derived affinity fingerprints (part 2): modeling performance for potency prediction

Fig. 5

Verification of the linear model assumptions. a Assumption of homoscedasticity of the residuals. Fitted values against the residuals. The residuals are centered around zero and, roughly, present a comparable dispersion across the range of values considered, indicating that the assumption of the homoscedasticity of the residuals is fulfilled. Assumption of the normality of the residuals, assessed with the distribution of the residuals (b) and a quantile–quantile (Q–Q) plot (c). The residuals follow a Gaussian distribution with zero mean, indicating that the assumption of the normality of the residuals is fulfilled. Overall, these results indicate that the assumptions of the linear model are fulfilled

Back to article page