pdig.0000176.g002.tif (774.13 kB)
Download fileChronological age prediction performance.
figure
posted on 2023-01-24, 18:34 authored by Alan Le Goallec, Sasha Collin, M’Hamed Jabri, Samuel Diai, Théo Vincent, Chirag J. PatelA.) R2 (R-squared) and B.) Mean Absolute Error (MAE). Model name contain the duration (e.g., full week), data type (e.g., Scalars or Time Series), and model algorithm type (e.g., LightGBM). * represents ensemble model. Red point is the overall ensemble across all models, green points denote the measured physical activity for the full week as inputs, and blue points denote the walking or gait data as inputs.
History
Usage metrics
Read the peer-reviewed publication
Categories
Keywords
mean absolute errormachine learning approachesgenome wide association7 ± 0271 scalar features113 time seriesraw frequency data456 one weekdefined accelerated agingaccelerated aging phenotypesg smoking ),div >< pderived biological agecomplex phenotype associated3 ± 0physical activity tends2 </ supaccelerated agingphysical activityvariables associatedone ’new phenotypedata structuresworld activityor5v1 </hist1h1c </uk biobankrelated diseasespredicted olderpredict ageolfactory clusterneural networkincreasing vulnerabilityg incomefour imagesclose proximitychromosome sixactual ageaccelerometer records9 %)