Selected article for: "area ROC curve and AUC ROC curve"

Author: Delli Ponti, Riccardo; Marti, Stefanie; Armaos, Alexandros; Tartaglia, Gian Gaetano
Title: A high-throughput approach to profile RNA structure
  • Document date: 2017_3_17
  • ID: k23xlzj0_24
    Snippet: From low-(top and bottom 50% of the CROSS score distribution) to high-confidence (top and bottom 5%) predictions, we observed an increase in the accuracies of our models, which indicates good ability to capture strong-signal regions. For instance, the accuracy of the icSHAPE-Mouse model applied to the SHAPE-HIV data set improves from 0.63 (low-confidence) to 0.81 (high-confidence; Figure 1B) , and the same trend is found with respect to other dat.....
    Document: From low-(top and bottom 50% of the CROSS score distribution) to high-confidence (top and bottom 5%) predictions, we observed an increase in the accuracies of our models, which indicates good ability to capture strong-signal regions. For instance, the accuracy of the icSHAPE-Mouse model applied to the SHAPE-HIV data set improves from 0.63 (low-confidence) to 0.81 (high-confidence; Figure 1B) , and the same trend is found with respect to other data sets (Supplementary Figures S4 and S5 We observed comparable cross-validation performances on the PARS datasets (area under the ROC curve AUC of 0.89 for PARS-Yeast applied to PARS-Human, and 0.90 for PARS-Human applied to PARS-Yeast), even though the experiments were carried out in different organisms and with slightly modified protocols, confirming the high quality of our predictions (Figures 2 and 3) .

    Search related documents:
    Co phrase search for related documents