A1 Refereed original research article in a scientific journal
Locality kernels for sequential data and their applications to parse ranking
Authors: Tsivtsivadze E, Pahikkala T, Boberg J, Salakoski T
Publisher: SPRINGER
Publication year: 2009
Journal: Applied Intelligence
Journal name in source: APPLIED INTELLIGENCE
Journal acronym: APPL INTELL
Volume: 31
Issue: 1
First page : 81
Last page: 88
Number of pages: 8
ISSN: 0924-669X
DOI: https://doi.org/10.1007/s10489-008-0114-2
Abstract
We propose a framework for constructing kernels that take advantage of local correlations in sequential data. The kernels designed using the proposed framework measure parse similarities locally, within a small window constructed around each matching feature. Furthermore, we propose to incorporate positional information inside the window and consider different ways to do this. We applied the kernels together with regularized least-squares (RLS) algorithm to the task of dependency parse ranking using the dataset containing parses obtained from a manually annotated biomedical corpus of 1100 sentences. Our experiments show that RLS with kernels incorporating positional information perform better than RLS with the baseline kernel functions. This performance gain is statistically significant.
We propose a framework for constructing kernels that take advantage of local correlations in sequential data. The kernels designed using the proposed framework measure parse similarities locally, within a small window constructed around each matching feature. Furthermore, we propose to incorporate positional information inside the window and consider different ways to do this. We applied the kernels together with regularized least-squares (RLS) algorithm to the task of dependency parse ranking using the dataset containing parses obtained from a manually annotated biomedical corpus of 1100 sentences. Our experiments show that RLS with kernels incorporating positional information perform better than RLS with the baseline kernel functions. This performance gain is statistically significant.
UTU Research Portal