In the model training phase, features selected using the Enet and Enet-subset methods were used to train two separate SVM models (SVM model #1 and SVM model #2, Figure 1D). As before, the features were normalized, and optimal model parameters were fed into each final SVM model. We used a grid-search algorithm to optimize the cost (C) of each SVM classifier. The search scale was set to C = 1:10, and the cost with the highest performance was used in each final model. To generalize the training process and obtain a more accurate model, we used a K-fold (K = 4) cross-validation, which was repeated five times. This technique divides data into equal disjointed subsets of size four. The model was then trained on all folds except one. The remaining subset was reserved for testing purposes. This process was then repeated three (K−1) times, selecting each fold to be used for testing once. We repeated this process five times to ensure that our trained model acquired most of the patterns from the training dataset.

We evaluated the performance of each SVM model using the test dataset where HCs were classified as positive and LBP as negative for the true positive (TP), false positive (FP), true negative (TN), and false negative (FN) calculations. We determined the corresponding accuracy, sensitivity, and specificity of each model. The accuracy (%) is defined as the fraction of correctly classified subjects {(TP+TN)/(TP+TN+FP+FN)}. Sensitivity is defined as the fraction of correctly classified positive samples from all positive samples {TP/(TP+FN)}. Specificity is defined as the fraction of correctly classified negative samples from all negative samples {TN/(TN+FP)}. We then determined the area under the receiver operating characteristics curve (AUC) to evaluate each model's overall performance.

Note: The content above has been extracted from a research article, so it may not display correctly.



Q&A
Please log in to submit your questions online.
Your question will be posted on the Bio-101 website. We will send your questions to the authors of this protocol and Bio-protocol community members who are experienced with this method. you will be informed using the email address associated with your Bio-protocol account.



We use cookies on this site to enhance your user experience. By using our website, you are agreeing to allow the storage of cookies on your computer.