Sandbox for non-linear signal processing in Python (2.7) written for a better understanding of the complex methods. All materiel and syntax is inspired by the work of Lars Kai Hansen DTU for the 02657 non-linear signal processing course.
- NumPy (http://www.numpy.org/) --
sudo apt-get install python-numpy
- Matplotlib (http://matplotlib.org/) --
sudo apt-get install python-matplotlib
cd ~ && git clone https://github.com/aaskov/nsp.git
This is a timeseries prediction example using a Gaussian Process which is found in example_gp.py
. Both the best log-likelihood and least-square fit is shown. A 95% confidence interval along the predictions is shown in the right figure.
This is a two-class classification example using a Support Vector Machine which is found in example_svm.py
. The algorihtm learns a set of support vectors that can be used for new (unseen) observations. The test result is shown in the right figure.
MIT