[SciPy-user] Combination of negative-exponential + normal distribution, how to tackle ?

[Stef Mientki]

hello,

Unfortunately statistics is not my strongest point,
so maybe someone can give me some hints or links,
with this problem.

I've recordings of the maximum activity of persons,
and I want to quantify these recordings.
Until now we just  calculated mean of the activity pattern.

If I look at some of the recordings,
(especially the ones of persons with high activities)
and split the recordings in week and week-end activity scores,
I think I can clearly see a combination of 2 distributions:

- person is in relative rest, leading to a negative exponential distribution
- person is active for a longer period of time, resulting in  normal 
distribution

For persons with a high activity,
I can estimate the estimate the mean and SD of the normal distribution,
simply by ignoring the low values, where the exponential distribution is 
high,
while the tail of the exponential distribution doesn't affect the normal 
distribution much.

But for persons with lower activity, the 2 distributions merge too much.
Is there a way to separate those distributions,
In fact I'm only interested in the parameters of the normal distribution.

thanks,
Stef Mientki



Kamer van Koophandel - handelsregister 41055629  / Netherlands Chamber of Commerce - trade register 41055629