hi
i did my phd research related to comparing neural network to svm. i managed to train svm to classify large number of feature inputs. the training time was way shorter than the neural network trainig time.
as a matter of fact, by increasing the number of features, data will become sparse, but it is def not an issue with svm or even neural network,
the difference between the two methods, is the relatively shorter time and the classification error were quite similar values.
exponential rbf, was a good method to classify the data.
kinds regards
Ramzi Fayad
---------------------------------------------------------------------------------------------
Ramzi Fayad
PhD researcher - Advanced Manufacturing Technology
Manufacturing Main Labs
Room B1,
School of Mechanical, Materials, Manufacturing
Engineering and Management
University of Nottingham
University Park
Nottingham
NG7 2RD
Phone: 0115 846 7939
Fax: 0115 351 4000
Email: [log in to unmask]
---------------------------------------------------------------------------------------------
>>> [log in to unmask] 02/16/05 11:16 PM >>>
Hello,
SVM is known to defeat the curse of dimensionality...then why is it so
that having a large number of attributes/features in the data is not
desirable when using SVM?
I thought it should not matter....where am I getting confused?
Thank You.
Sincerely,
Monika Ray
***********************************************************************
The sweetest songs are those that tell of our saddest thought...
Computational Intelligence Centre, Washington University St. louis, MO
**********************************************************************
This message has been checked for viruses but the contents of an attachment
may still contain software viruses, which could damage your computer system:
you are advised to perform your own checks. Email communications with the
University of Nottingham may be monitored as permitted by UK legislation.
|