Fig. 3From: High resolution MRI-based radiomic nomogram in predicting perineural invasion in rectal cancerFeature selection and dimensionality reduction. a 10-fold cross-validation of the LASSO analysis was performed to select the most valuable features in predicting PNI. The abscissa corresponding to the lowest point of model deviation is the optimal lambda value, that is, the position of the first dashed line. b The regression coefficients of LASSO. Each colored line represents the variation curve of the characteristic coefficient with the lambda value. The lambda value (the position represented by the dashed line) found in Fig. 3a is used to determine which parameter has a coefficient that is not 0, then this parameter is used in the final Model building. c The final features and corresponding coefficients. The blue bars show corresponding coefficients of each final feature, indicating the importance of PNI predictionBack to article page