Selected article for: "cross validation and random forest"
Author: Yujia Xiang; Quan Zou; Lilin Zhao
Title: VPTMdb: a viral post-translational modification database
  • Document date: 2020_4_2
    • ID: kl99afiu_65
    Snippet: The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/2020.04.01.019562 doi: bioRxiv preprint classifiers' performance from the 5-fold cross validation (Supplementary Table S3 ). Figure 3A shows that the maximum AUCs of the svm and random forest are similar. For the random forest and svm, the AUCs increased when more features were selected (random forest: 14-135 features, with AUC > 0.90; svm: 27-13.....
    Document: The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/2020.04.01.019562 doi: bioRxiv preprint classifiers' performance from the 5-fold cross validation (Supplementary Table S3 ). Figure 3A shows that the maximum AUCs of the svm and random forest are similar. For the random forest and svm, the AUCs increased when more features were selected (random forest: 14-135 features, with AUC > 0.90; svm: 27-135 features, with AUC > 0.90). However, we observed that the AUCs of naïve Bayes (AUCs > 0.80) decreased when more features were added. From a statistical point of view, to prevent the curse of dimensionality, fewer and more meaningful features should be chosen. Taking the above results into consideration, for 68 features, the AUCs of the three predictors perform better, suggesting that 68D is the most meaningful feature among all the features.

    Search related documents:
    Co phrase search for related documents
    • better perform and random forest: 1, 2, 3, 4, 5, 6, 7, 8
    • better perform and random svm forest: 1, 2
    • cross validation and feature select: 1
    • cross validation and maximum auc: 1
    • cross validation and predictor auc: 1
    • cross validation and random forest: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25
    • cross validation and random svm forest: 1, 2, 3, 4, 5, 6, 7
    • feature select and random forest: 1, 2, 3
    • feature select and random svm forest: 1
    • predictor auc and random forest: 1, 2