Selected article for: "sensitivity specificity and test set"

Author: Andre Filipe de Moraes Batista; Joao Luiz Miraglia; Thiago Henrique Rizzi Donato; Alexandre Dias Porto Chiavegatto Filho
Title: COVID-19 diagnosis prediction in emergency care patients: a machine learning approach
  • Document date: 2020_4_7
  • ID: nvavj9gk_12
    Snippet: The two algorithms with best overall performance were random forests and support vector machines with an AUC of 0.85 for the test set. Figure 1 presents the AUC of the five algorithms, which shows that the performance was similar between the algorithms with some overlap across the different discriminative thresholds. The two algorithms with the best predictive performance (support vector machines and random forests) had the same discrimination re.....
    Document: The two algorithms with best overall performance were random forests and support vector machines with an AUC of 0.85 for the test set. Figure 1 presents the AUC of the five algorithms, which shows that the performance was similar between the algorithms with some overlap across the different discriminative thresholds. The two algorithms with the best predictive performance (support vector machines and random forests) had the same discrimination results (sensitivity of 0.677 and specificity of 0.850), but the support vector machines algorithm had a slightly better calibration, with a Brier score of 0.160 (Table 2 ). Every one of the five algorithm had a positive predictive value of at least 0.74 and a negative predictive value of at least 0.77. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

    Search related documents:
    Co phrase search for related documents