File:PGSPredictionPerformance VS sampleSize RabenLelloEtAl.svg
![File:PGSPredictionPerformance VS sampleSize RabenLelloEtAl.svg](http://upload.wikimedia.org/wikipedia/en/thumb/2/26/PGSPredictionPerformance_VS_sampleSize_RabenLelloEtAl.svg/546px-PGSPredictionPerformance_VS_sampleSize_RabenLelloEtAl.svg.png?20231117055812)
Original file (SVG file, nominally 546 × 335 pixels, file size: 58 KB)
✓
The source code of this SVG is valid.
Summary
[edit]Description |
PGS predictor performance increases with the dataset sample size available for training. Here illustrated for hypertension, hypothyroidism and type 2 diabetes. The x-axis labels number of cases (i.e. samples with the disease) present in the training data and uses a logarithmic scale. The entire range is from 1,000 cases up to over 100,000 cases. The numbers of controls (i.e. samples without the disease) in the training data were much larger than the numbers of cases. These particular predictors were trained using the LASSO algorithm. |
---|---|
Source |
adapted with permission from https://arxiv.org/abs/2101.05870 |
Date |
2021-01-14 |
Author |
Tim G. Raben, Louis Lello, Erik Widen and Stephen D.H. Hsu |
Permission (Reusing this file) |
CC BY 4.0 https://creativecommons.org/licenses/by/4.0/
|
Licensing
[edit]![]() ![]() | This work is licensed under the Creative Commons Attribution 4.0 License. |
File history
Click on a date/time to view the file as it appeared at that time.
Date/Time | Thumbnail | Dimensions | User | Comment | |
---|---|---|---|---|---|
current | 05:58, 17 November 2023 | ![]() | 546 × 335 (58 KB) | Minorax (talk | contribs) | fix // Editing SVG source code using c:User:Rillke/SVGedit.js |
15:32, 14 April 2021 | ![]() | 546 × 335 (60 KB) | Stal potaten (talk | contribs) | {{Information |Description = PGS predictor performance increases with the dataset sample size available for training. Here illustrated for hypertension, hypothyroidism and type 2 diabetes. The x-axis labels number of cases (i.e. samples with the disease) present in the training data and uses a logarithmic scale. The entire range is from 1,000 cases up to over 100,000 cases. The numbers of controls (i.e. samples without the disease) in the training data were much larger than the numbers of cas... |
You cannot overwrite this file.