Large Scale Learning with String Kernels

In applications of bioinformatics and text processing, such as splice site recognition and spam detection, large amounts of training sequences are available and needed to achieve sufficiently high pre