biobankAccelerometerAnalysis

biobankAccelerometerAnalysis copied to clipboard

https://github.com/activityMonitoring/biobankAccelerometerAnalysis/blob/0b3879386b06fc4bd41714b91de1ba59c34b892d/accelerometer/summariseEpoch.py#L275-L280 Currently `quality-goodWearTime` only supports hard-coded threshold which is not ideal.

angerhang

Check gt3x v1 and v2 timestamp differences

1

Thanks @chanshing for bringing this up. > Just to double check, is this line correct?: >https://github.com/activityMonitoring/biobankAccelerometerAnalysis/blame/6021de453950e464737373cd62ea7c6a6af3a391/java/ActigraphReader.java#L479 > Also, I see that in some parts firstSampleTime is multplied by 1000 to...

angerhang

1. Get the sample files for all the supported accelerometer file types 2. Generate the epoch files as reference 3. Create a testing workflow

angerhang

As Shing pointed out, our current way of calculating the correlation might suffer from catastrophic cancellation. Good to update to a more stable algorithm for covariance. Even though we don't...

angerhang

Bumps [joblib](https://github.com/joblib/joblib) from 1.1.0 to 1.2.0. Changelog Sourced from joblib's changelog. Release 1.2.0 Fix a security issue where eval(pre_dispatch) could potentially run arbitrary code. Now only basic numerics are supported....

dependabot[bot]

https://biobankaccanalysis.readthedocs.io/en/latest/usage.html#training-a-bespoke-model

chanshing

chanshing

@aidanacquah @LauraBrocks

chanshing

Extract light from ax3

5

* Include lux extraction from ax3 to epoch file * Add dummy light variable to device readers without light * Include summary stat calculation for light Test plan: ``` python...

angerhang

The feature set used to train the released versions of models was slightly different to that used in the original publications. Ideally this would be documented somewhere.

R-Walmsley