Performance gains in activity flow prediction after combining spatial and functional embedding of routes.
<p>Accuracy of activity flow prediction based on FC <b>(A)</b>, FC(SE) <b>(B)</b>, and FC(FE, SE) <b>(</b><b>C</b><b>)</b> for different datasets. Accuracy of activity flow prediction based on SPL <b>(D)</b>, SPL<sub>...
محفوظ في:
| المؤلف الرئيسي: | |
|---|---|
| مؤلفون آخرون: | , , , , , |
| منشور في: |
2025
|
| الموضوعات: | |
| الوسوم: |
إضافة وسم
لا توجد وسوم, كن أول من يضع وسما على هذه التسجيلة!
|
| الملخص: | <p>Accuracy of activity flow prediction based on FC <b>(A)</b>, FC(SE) <b>(B)</b>, and FC(FE, SE) <b>(</b><b>C</b><b>)</b> for different datasets. Accuracy of activity flow prediction based on SPL <b>(D)</b>, SPL<sub>wei</sub>(SE) <b>(E)</b>, and SPL<sub>wei</sub>(FE, SE) <b>(</b><b>F</b><b>)</b> for different datasets. Accuracy of activity flow prediction based on SPL <b>(G)</b>, SPL<sub>bin</sub>(SE) <b>(H)</b>, and SPL<sub>bin</sub>(FE, SE) <b>(</b><b>I</b><b>)</b> for different datasets. Statistical significance was identified based on the area under the curve (AUC) across all density thresholds. FC, functional connectivity; FE, functional embedding; SE, spatial embedding; SPL<sub>wei</sub>, shortest path length based on weighted network; SPL<sub>bin</sub>, shortest path length based on binary network; ns, nonsignificant. *<i>p</i> < 0.05, ***<i>p</i> < 0.001 (<i>p</i> < 0.05, Bonferroni corrected).</p> |
|---|