Abstract
Dark adaptation can be modeled using a seven parameter composite function. I measured this function in 35 subjects. An algorithm was used that determined the parameters for each subject. There is no theoretical route to determine confidence intervals for these parameters. Therefore a bootstrap method using the high through-put computer cluster was used. This reduced the total calculation time from 14 days to 5 hours. MSE is not a useful substitute for bootstrapped confidence intervals.
Original language | English |
---|---|
Publication status | Published - 2011 |
Event | Faculty of Life Sciences Symposium - Armitage Centre, Manchester, United Kingdom Duration: 23 Sept 2011 → 23 Sept 2011 |
Keywords
- Dark adaptation, nonlinear, confidence intervals, Poster