Kernel methods are powerful tools in machine learning. They have to be computationally efficient. In this paper, we present a novel list-based approach to compute efficiently the string subsequence kernel (SSK). Our main idea is that our list-based SSK reduces to range query problem. We started by the construction of a match list L(s,t)={(i,j):si = tj } where s and t are the strings to be compared; such match list contains only the required data that contribute to the result. To do some intermediate processing efficiently, we constructed a layered range tree and applied the corresponding computational geometry algorithms. Moreover, we extended our match list to be a list of lists in order to improve the computation efficiency of the SSK. The whole process takes O(|L|log|L|+pK) time and O(|L|log|L|+K) space, where |L| is the size of the match list, p is the length of the SSK and K is the total reported points by range queries over all the entries of the list. © 2014 Springer International Publishing.
CITATION STYLE
Bellaouar, S., Cherroun, H., & Ziadi, D. (2014). Efficient list-based computation of the string subsequence kernel. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 8370 LNCS, pp. 138–148). Springer Verlag. https://doi.org/10.1007/978-3-319-04921-2_11
Mendeley helps you to discover research relevant for your work.