A significant trend to higher AUC was also observed in the group with 10-fold validation features [80.1% (95% CI: 78.1%–82.1%), with a prediction interval of 72.4%–87.8%] compared to fivefold validation features [81.5% (95% CI: 80.0%–83.1%), with a prediction interval of 74.0%–89.1%] ( P < 0.001) in the analysis of all models; however, the results for the best models were inconclusive and not statistically significant ( P = 0.703).
← all excerpts
Role of artificial intelligence in developing predictive models for major adverse cardiovascular outcomes using CCTA adipose tissue characteristics: a systematic review and meta-analysis.
1
—
—