0

I have 2 data vector (data1 in blue, data2 in red) representing some sort of probabilities (they are not really probability, i.e. sum could be different from 1). As you can see from the picture below data1 could be approximated with a Gaussian distribution.

data2 should follow a bivariate gaussian distribution (only the peaks close to the start and end of the blue line should be considered, the rest it could be considered as noise).

My question is: Is there a way in matlab to approximate the blue and red line with a gaussian and bivaraite gaussian distribution? How?

I would like to do so to filter out ther first bump in data1 and the first and the 3rd bumps in data2. I would like to use this approach instead of some smoothing...

Here you can download the data1 and data2 to reproduce the graph.

Thanks

enter image description here

gabboshow
  • 5,359
  • 12
  • 48
  • 98
  • 2
    It strikes me that it would be impossible to filter out the first and third peaks of data2 - the first peak is larger than the fourth! You'd need to give a lot more information about the background distribution before you could hope to see any success... – srthompers Jul 13 '15 at 16:49
  • Hi, thanks for your comment. Actually this is a secondary problem for me...once I have a nice gaussian distribution for data1 I can take the closest peaks to the start of the rising and falling part of the gaussian.. – gabboshow Jul 13 '15 at 16:52

0 Answers0