37 lines
1012 B
Plaintext
37 lines
1012 B
Plaintext
|
Notable changes from previous release:
|
||
|
|
|
||
|
|
In development:
|
||
|
|
|
||
|
|
Version 1.4.1:
|
||
|
|
* Fix bug in CEEMDAN SNR fixing introduced in 1.4; in some cases
|
||
|
|
(especially N=16) the noise can have zero variance causing an overflow
|
||
|
|
|
||
|
|
Version 1.4:
|
||
|
|
* pyeemd splitted to a separate project
|
||
|
|
* Easier installation on OSX
|
||
|
|
* The GSL error handler is disabled and realistic errors are instead
|
||
|
|
handled by libeemd
|
||
|
|
* In CEEMDAN the SNR is fixed separately for each mode as done by Torres
|
||
|
|
et al
|
||
|
|
|
||
|
|
Version 1.3.1:
|
||
|
|
* Minor documentation improvements
|
||
|
|
* Minor bug fix about possible redefinition of NDEBUG
|
||
|
|
|
||
|
|
Version 1.3:
|
||
|
|
* New default stopping parameters for pyeemd
|
||
|
|
* pyeemd warns if using num_siftings=0
|
||
|
|
* Completely reworked S-number criterion handling
|
||
|
|
* Minor bug fixes
|
||
|
|
* Minor improvements to unit tests
|
||
|
|
|
||
|
|
Version 1.2:
|
||
|
|
* Rudimentary error reporting in the C API
|
||
|
|
* Allow setting number of IMFs to compute
|
||
|
|
|
||
|
|
Version 1.1:
|
||
|
|
* Implementation of CEEMDAN
|
||
|
|
* Allow setting the RNG seed
|
||
|
|
* Documentation improvements
|
||
|
|
* Minor optimizations
|