The ISMIR2004 Music Genre:
http://ismir2004.ismir.net/genre_contest/index.htm
(0)Split the data into 90% and 10% using weka:
http://weka.wikispaces.com/How+do+I+divide+a+dataset+into+training+and+test+set%3F
(1)instead of manually produce the mf and arff file, we can use the filter in weka to do this: weka.filters.supervised.instance.StratifiedRemoveFolds -S 0 -N 10 -F 1
(2) combine two arff file into one:
Usage:
weka.core.Instances help
weka.core.Instances <filename>
weka.core.Instances merge <filename1> <filename2>
weka.core.Instances append <filename1> <filename2>
weka.core.Instances headers <filename1> <filename2>
weka.core.Instances randomize <seed> <filename>
(3) Add instance ID http://weka.wikispaces.com/Instance+ID
(4) Change the visilization effect http://weka.wikispaces.com/weka_gui_visualize_Visualize.props