Dmitry: This is really fine, fine work. At the moment i am using your App for a professional project of mine in which i am working on a model to predict the outcome in US court decisions. I've found it to be remarkably stable and also very fast.
Kind Regards,
--doug
Name
indra
Date
2008-05-29
Location
Message
hai
Name
Emanuele Olivetti
Date
2008-03-11
Location
Message
Me and my research group are using OpenOpt in Machine Learning. The task is model selection. The nice architecture of this library and its powerful algorithms allowed us to be much more effective in our work.
Since we develop our own algorithms in Python and C code, it is extremely easy for us to use this library.
We hope the best for the OpenOpt project.
Emanuele http :// nilab . fbk . eu
Name
Mclean Edwards
Date
2008-01-30
Location
Message
Openopt (along with GenericOpt) is a wonderful concept brought to life. I was going to start something similar to aid me in my PhD research when I discovered this project. There is no way I would have had enough time to accomplish what Dmitrey has in the past year.
Many optimization algorithms exist in code, but are spread across languages and libraries that the easiest way to explore and compare the algorithms has been to recode the majority of them.
Now is the time for convergence into a unified framework for learning, research, and application. Python is a good language to use, SciPy is an excellent collection of tools, and openopt on top of these resources appears to be the best location for this convergence to happen.
I look forward to working closely with this package and encouraging its use within the optimization community in Canada and beyond.
Dmitrey has done a good job here, and his work should be recognized. I especially appreciate the online presentation from the Newsline to the Source Tracker and Documentation.
Name
Great Zamfir
Date
2008-01-14
Location
Message
Hi, I am currently working on an aerodynamic optimization system for my Master's thesis. I'm using scipy and your optimization framework. At the moment I am mainly using the Powell method and Nelder-Mead already incorporated in Scipy, not your new algorithms, but I still like the framework.