Skip to main content
. 2024 Nov 15;6(6):fcae393. doi: 10.1093/braincomms/fcae393

Figure 1.

Figure 1

Hyperparameter grid search of the FS and ML algorithm combinations by using training data with 5-fold cross-validation. (A–C) The AUC results of FS and ML algorithm combinations based on clinical + radiological + radiomics variables with (A) mask DWI, (B) mask ADC620 and (C) mask ADC. (D) The AUC results of ML algorithm based on clinical + radiological variables with the above three types of masks. Adaboost, adaptive boosting; CIFE, common and individual feature extraction; CMIM, conditional mutual information maximization; DET, deep extremely randomized trees; DISR, dental image segmentation and retrieval; EXT, extremely randomized trees; Fast ICA, fast independent component analysis; ICAP, interaction capping; JMI, joint mutual information; KNN, K-nearest neighbour; MIM, mutual information maximization; MLP, multi-layer perceptron; NMF, non-negative matrix factorization; None, all features without any selection; PCA, principal component analysis; Truncated SVD, truncated singular value decomposition; Xgboost, extreme gradient boosting.