Fast kernels for string and tree matching
Vishy Vishwanathan and Alex Smola
Kernels and Bioinformatics
, Cambridge, MA, 2004
In this chapter we present a new algorithm suitable for matching discrete objects
such as strings and trees in linear time, thus obviating dynamic programming with
quadratic time complexity.
This algorithm can be extended in various ways to provide linear time prediction
cost in the length of the sequence to be classified. We demonstrate extensions in the
case of position dependent weights, sliding window classifiers for a long sequence,
and efficient algorithms for dealing with weights given in the form of dictionaries.
This improvement on the currently available algorithms makes string kernels a
viable alternative for the practitioner.