I am planing to write paper about detecting sleep apnea with large volume data and i have submitted my data use proposal
I ask you several point about my wondering
If the model trained with several none-commercial purpose data set shows good performance our commercial-used device data, can i conclude that "this result represent that our proposed method decrease development cost." if it is not possible, should i conclude that "this result represent that our proposed method decrease development cost if these datasets are non-commercial use datasets"?
thank for your reply !
In my research which estimates obstruct sleep apnea severity with ECG, i found that it is possible to estimate obstruct sleep apnea severity correctly from other commercial device ECG using a model feeded with CFS and MESA ECG.
So i think that it can decrease commercial device development cost and i has one question that there are any problem if these data are commercially used.
If you could take a few minutes to answer this questions, i would really appreciate it.
I have confirmed that there is no problem with 2852 subject of mesa data ("slpprdp5" variable --> 393, the TST from new xml files sleep stage --> 393)
Thanks for your kindly reply
I has totaly understanded your answers and written down two data requirement forms (cfs, mesa) you asked me
I will re-download data and review one more time
Thanks for your kindly reply,
I send you information you requested
. about "variable TST"
CFS: "SLPPRDP" variable
MESA: "slpprdp5" variable
. account issue
Yes i has obtained access to data on a different account, one of the our team member got access for data.
If you need to a information about the account, i will inform you that
HI, i am on sleep study with CFS & MESA data
on the study i has found critical issue can affect other future works
this issue should be identified promptly.
you know, total sleep time (TST) is can be calculated from sleep stage like equation 1 and the sleep stage can be extracted from annotation xml file
equation1) TST = in bed time - (wake after sleep onset + sleep latency)
Common sense, a TST from the sleep stage and TST from variable csv file are must be same, however several subject are difference
To check the exact value of errors, i ploted figures can express the two values.
i has attached the plots and writed down details, i hope you to deal with this issue promptly
plot figures link
MESA (1899 subjects), unit: min
(subject number, from variable TST, from sleep stage TST)
(1738, 226, 256)
(2852 , 393, 502.5)
(6476 , 403, 411)
CFS (694 subjects), unit: min
(800937, 19, 28)
(800170, 248, 265)
(801121, 269, 276.5)
(802726, 276, 283.5)
(802368, 326, 340)
(802678, 335, 362.5)
(802001, 359, 363.5)
(801376, 380, 383.5)
(800402, 379, 432.5)