I reached a great milestone with SVMMail today. I will be doing more test and releasing more information next week. The initial results are a 97% accuracy on the filter. Also with real world testing (so far a low number of 65 mails), there was only 2 errors (1 false positive) in prediction. I had training data of 550 real emails and 713 spam emails (all of which i collected in the 3 weeks or so that I ahve been working on this project.) I am really excited that I have past the stumbling blocks that I was on the last 3 days where I was actually getting a 0% accuracy because a bug the was generating a pretty much random model.
There is currently no web interface and it is all just run directly from java (jbuilder in my case). I will add features like that and the ability to track how many of each type of error my system makes later.
This is a great I am really happy with how this is working out.
Comments (2)
Dear sir,
I am a student of university of mauritius.i am presently working on a project concerning email authorship analysis and according to my research i have to use an SVM machine for this. I am very new to this field. I have found out that for using the SVM. I have to get a dataset file from my inbox. Can u explain to me the procedures for using an SVM machine.
thank you in advance.
ti
Posted by ti | June 2, 2004 6:32 AM
Posted on June 2, 2004 06:32
Getting all the messages from your inbox is quite easy and seperate from using the SVM machine. First you should look on my java help page because it contains links to two sites all about email retrival. Then if you want to learn how to format the email text to work with SVM you should look at my text2svm tool. It shouldn't be to hard to use. You should also download SVM light and do the tutorials with the available datasets so you know how to use the tool and are aware of how the results come out. Hopefully this helps if your having a more specific problem you can email me or leave me a comment here.
Posted by Dan Mayer | June 10, 2004 6:17 AM
Posted on June 10, 2004 06:17