A lot of the coverage of this paper has focused on the machine learning piece of this project. I am glad because I think that is actually the most interesting part of the project. The idea came from the way I listen to MP3's. I don't like playlists, so I just put my player on random and skip songs that I don't want to hear. This whole project is really a way for me not to have create playlists.
The neural network component of XPod did not perform as well as I would have liked. I am now replacing the neural network with a support vector machine. SVM's are the cool new kid on the machine learning block. I don't know much about them, but they seem to out perform almost every other algorithm. XPod will be reworked for my machine learning class. This class has a good track record for success. Last semester a student created a very successful sblog detection scheme that is now being perused by major blog search engines.
The neural network component of XPod did not perform as well as I would have liked. I am now replacing the neural network with a support vector machine. SVM's are the cool new kid on the machine learning block. I don't know much about them, but they seem to out perform almost every other algorithm. XPod will be reworked for my machine learning class. This class has a good track record for success. Last semester a student created a very successful sblog detection scheme that is now being perused by major blog search engines.
posted by Sandor Dornbush at
10:02 AM
|
0 comments
