@user-dc89dc porting the conversation from core channel here regarding cognitive load.

Have you taken a look at other research papers - https://docs.google.com/spreadsheets/d/1ZD6HDbjzrtRNB4VB0b7GFMaXVGKZYeI0zBOBEEPwvBI/edit?usp=sharing

Hi @wrp thanks for the link!

I've seen a couple of old papers that feature engineer on gaze data, to train a classifier for cognitive load

such as Liang et al., 2007 - https://ieeexplore.ieee.org/document/4220657/

Anyone here work on feature engineering?

Mine looks something like: `{avg_position, stddev_position, avg_speed, stddev_speed, [email removed] [email removed] etc

These statistics are calculated at every gaze sample over the 5 preceding seconds of data

(Essentially rolling windows)

Wondering if anyone has seen good predictive performance from other possible rolling features?

Don't think any of the pupil papers have done this so far(!?)

@user-41f1bf bit of an old thread, but have you checked out GazeNET? https://www.researchgate.net/profile/Raimondas_Zemblys2/publication/319503441_End-to-end_eye-movement_event_detection_using_deep_neural_networks/links/59afa1f6458515150e496cac/End-to-end-eye-movement-event-detection-using-deep-neural-networks.pdf

Thank you for that link @user-dc89dc

Hey guys, does anyone have any experience with calculating gaze transition entropy and stationary entropy? I have found relative research works by KRZYSZTOF KREJTZ et al. - Gaze Transition Entropy but I am finding it difficult to implement the process discussed. Anything that I should take into consideration? Thanks in advance.