String Kernels Based on Variable-Length-Don't-Care Patterns

Abstract. We propose a new string kernel based on variable-lengthdon't-care patterns (VLDC patterns). A VLDC pattern is an element of ({}) , where is an alphabet and is the variable-length-don't-care symbol that matches any string in . The number of VLDC patterns matching a given string s of length n is O(22n ). We present an O(n5 ) algorithm for computing the kernel value. We also propose variations of the kernel which modify the relative weights of each pattern. We evaluate our kernels using a support vector machine to classify spam data.

• Type: Conference
• Year: 2008
• Where: DIS
• Authors: Kazuyuki Narisawa, Hideo Bannai, Kohei Hatano, Shunsuke Inenaga, Masayuki Takeda
• Keywords: DIS 2008, Kernel Value, String, Theoretical Computer Science, Vldc Patterns