Performances for regression models showed a tradeoff where better training led to worse testing and vice versa. Results: Area under the curve (AUC) performances varied considerably across the different data splits (e.g., radiomics regression model: train 0.58-0.70, test 0.59–0.73). For each split and classifier type, multiple models were created based on radiomics and/or clinical features. Logistic regression with regularization and support vector machine were used as the machine learning classifiers. For each split, cross-validation was used for training, followed by an assessment of the test set. The dataset was repeatedly shuffled and split into training (n=400) and test cases (n=300) forty times. Methods: Mammograms from 700 women were used to study upstaging of ductal carcinoma in situ. Objectives: To assess the performance bias caused by sampling data into training and test sets in a mammography radiomics study.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |