Releases: cvjena/libmaxdiv
Releases · cvjena/libmaxdiv
libmaxdiv v1.1
This minor release mainly comprises bugfixes and a few usability improvements.
libmaxdiv
- Added a function for computing the anomaly score of given individual intervals
- Fixed index vector comparison in
DataTensor::crop
(#3) - Resolved some compiler warnings
Python interface
- Added a function for computing the anomaly score of given individual intervals
- Do not import 3rd-party packages needed by the Python implementation of MDI if the C++ library is used
GUI
- Improving plotting/panning/zooming speed
- Fixed sometimes broken zooming
- Explicitly set linewidth instead of depending on the default of the installed matplotlib version
- Homogenous background color on Linux
libmaxdiv v1.0
Initial release of libmaxdiv.
Features:
- Divergences: Kullback-Leibler, Jensen-Shannon, Cross Entropy
- Distribution models: KDE, Gaussian, ERPH
C++
library withC
-style interface and Python bindings- Interval Proposals for high-speed processing
- Support for spatio-temporal data
- Support for data with missing values
- Deseasonalization
- Graphical User Interface