GitLab

computes the weights and the nodes for approximating gaussian distributions using the scaled unscented approach.

computes the proposal distribution during the importance sampling step for gaussian-mixture filters. Fix many bugs.

gaussian mixture nonlinear filters: uses gaussian-mixture approximations for particles. Fix bugs and normalize the way likelihood is calculated.

computes the proposal distribution during the importance sampling step in gaussian nonlinear filters. Uses a nonlinear Kalman filter and several gaussian approximations.

fix bugs and normalize the way we write the likelihood.

test different cases before doing resampling

fix bug for calculating observed predicted mean and variance with the correct weights.

-add a variable to make a difference between gaussian approximations during importance sampling and approximation of states (for gaussian and gaussian-mixture filters)
-modify the default of parameters for the gaussian approximation using the scaled unscented transformation

Added a unitary test in the overloaded colon method of the dynDate class, using the new feature introduced in the previous commits.

Changed subsref overloaded method (dynDate): added a new interface to populate empty dynDate objects with specified frequency. Added unitary test.

For instance if we instantiate an empty dynDate object for quaterly dates:

qq = dynDate('Q');

Then we can populate this object as follows (2009 second quarter):

d1 = qq(2009,2);

Added the possibility to instantiate empty quaterly, monthly or weekly dynDate objects. Added unitary test.