SDE Toolbox is a MATLAB package for simulating sample paths of the solution of a Ito or Stratonovich stochastic differential equation (SDE), estimate parameters from data and
i) this version includes faster implementations of the parameter estimation procedures:
the parametric estimation procedure (SDE_PSML.m) speed has been boosted for the case
of multi-dimensional SDEs, now it is 14x-27x times faster, depending on the machine; negligible
improvement for the non-parametric estimation procedure (SDE_NPSML.m); no improvement for
the estimation of one-dimensional SDEs.... read more
SDE TOOLBOX v. 1.4.0 released
"SDE Toolbox" is a MATLAB package for simulating sample paths of the solution of a Ito or
Stratonovich stochastic differential equation (SDE), estimate parameters from data and
Major new features:
1)with the new version (1.4.0, available at http://sourceforge.net/projects/sdetoolbox/\) the
use of global variables is avoided (except for the demo files): this makes the present version
much more capable of interfacing with other Matlab programs, but the structure of the 'sdefiles'
is different. Thus sdefiles created under previous versions will not work with v. 1.4.0
(however, it is straightforward to adapt them, just look at the new sdefiles in the "models_library"
folder); as a consequence the Toolbox can be used without necessarily running SDE_library_run.m, and
can be integrated into user defined Matlab programs; several examples are provided in the User's Guide
2) approximated parameters 95% confidence intervals can now be calculated. ... read more
SDE TOOLBOX - AN INTRODUCTION TO THE SIMULATION AND THE NUMERICAL SOLUTION OF STOCHASTIC DIFFERENTIAL EQUATIONS WITH MATLAB
"SDE Toolbox" is a MATLAB package for simulating sample paths of the solution of a Itô or Stratonovich stochastic differential equation (SDE), estimate parameters from data and visualize statistics. Users may simulate an SDE chosen from the provided model-library or implement their own one.... read more
There was a bug in the toolbox sdefiles' code: it only affected multi-dimensional SDE (not one-dimensional): thus simulations/results obtained with the SDE_Toolbox 1.2 (and previous versions) are correct if a one-dimensional SDE has been considered, otherwise results may be wrong. In version 1.2.1 all the sdefile were rewritten so, if you wrote your own sdefiles by modifying the files in the SDE_library, you should modify your codes to follow the new versions. However, as stated above, that is only necessary for d-dimensional SDE (d > 1). Sorry for the inconvenience.... read more
A major bug has been discovered into the SDE Toolbox: this bug should only affect multidimensional (not one-dimemensional) SDE model. However, a fixed version will be uploaded shortly.
Sorry for the inconvenience,
I am working on the next release of SDE Toolbox: in particular I want to include parameter estimation tools, i.e. given a set of data the package should iteratively compute approximated maximum likelihood estimates. I am planning to consider a parametric and a non-parametric method however, to see wheter I will be able to effectively include these tools I first have to check IF my MATLAB package will be fast enough to accomplish the task (I do not want to include MEX files, in order to maintain the code portable). This is a very difficult task so...let's hope for the best!... read more
Here is the project webpage:
In this version, for each SDE model Monte-Carlo statistics are displayed (moments, variance, 95% confidence levels, skewness, kurtosis etc.) as well as histograms of the process distribution at the endpoint T. The user's guide has been enalarged with topics on Wiener process simulation and a section describing the statistics above.
I have just uploaded the first version (...actually v. 1.1) of my little contribution. As far as I know there are few projects in the world devoted to SDE modeling, so I hope most of you will appreciate my work. Of course it needs much more work to be really useful but anyway, someone may still be interested in it.
ps: if you want to know something more on my researches, take a look at http://www.biomatematica.it/Pages/Picchini.html