Nonlinear growth models: An alternative to ANOVA in tomato trials evaluation