Prev: error with intial conditions for ode45
Next: PSO for image enhancement
From: MBALOVER on 12 Mar 2010 22:29
Hi all,

In my class project, I need to develop a classifier based on Logistic
Regression to classify ham and spam emails.

The request is that I use conjugate gradient algorithm to estimate a
parameter vector theta (that is the weights of words in emails). The
problem is that dimension of the vector theta is equal the number of
words in the training set that is a huge number (150,000).

I am wondering if it is possible to use MATLAB optimization tool box
for problems with variable of 150,000 dimensions.

If it is impossible, is there any free optimization tool that I can
use for this problem?

Thanks a lot.
 | 
Pages: 1
Prev: error with intial conditions for ode45
Next: PSO for image enhancement