Message Boards Message Boards

2
|
5468 Views
|
0 Replies
|
2 Total Likes
View groups...
Share
Share this post:

Mining Time Series Data. Wavelet transform

Posted 9 years ago

I'm working with epidemiology time series of tick-borne encephalitis. I want to extract some latent information about signal using wavelet transform (WT). As i understood from different articles and books information i can "take" from scalogram, but there are continuous and discrete WT and many type of "mother" wavelet. Can you help me to make choice (maybe some criteria of choosing) and tell me please what else i can do with signal to extract more information using wavelet?

Now i've got pictures like this: (Left - time series of infected people from weeks (vertical lines is years); Central - wavelet scalogram (frequency(weeks^-1) from weeks; continuous wavelet transform with gabor wavelet); Right - 3D scalogram) this is data for 6 regions since 2010 to 2014 years

Draft version of my calcus i add to this post.

enter image description here

enter image description here

Attachments:
POSTED BY: Alex Skudarev
Reply to this discussion
Community posts can be styled and formatted using the Markdown syntax.
Reply Preview
Attachments
Remove
or Discard

Group Abstract Group Abstract