tech,4-1-P03-1005,ak results </term> . This paper proposes the <term> Hierarchical Directed Acyclic Graph ( HDAG ) Kernel </term> for <term> structured natural language
other,13-1-P03-1005,ak Acyclic Graph ( HDAG ) Kernel </term> for <term> structured natural language data </term> . The <term> HDAG Kernel </term> directly
tech,1-2-P03-1005,ak natural language data </term> . The <term> HDAG Kernel </term> directly accepts several levels of
other,9-2-P03-1005,ak directly accepts several levels of both <term> chunks </term> and their <term> relations </term> ,
other,12-2-P03-1005,ak of both <term> chunks </term> and their <term> relations </term> , and then efficiently computes the
measure(ment),19-2-P03-1005,ak and then efficiently computes the <term> weighed sum </term> of the number of common <term> attribute
other,26-2-P03-1005,ak sum </term> of the number of common <term> attribute sequences </term> of the <term> HDAGs </term> . We applied
tech,30-2-P03-1005,ak <term> attribute sequences </term> of the <term> HDAGs </term> . We applied the proposed method
tech,6-3-P03-1005,ak We applied the proposed method to <term> question classification </term> and <term> sentence alignment tasks
tech,9-3-P03-1005,ak <term> question classification </term> and <term> sentence alignment tasks </term> to evaluate its performance as a <term>
measure(ment),18-3-P03-1005,ak </term> to evaluate its performance as a <term> similarity measure </term> and a <term> kernel function </term>
tech,22-3-P03-1005,ak <term> similarity measure </term> and a <term> kernel function </term> . The results of the experiments
tech,8-4-P03-1005,ak the experiments demonstrate that the <term> HDAG Kernel </term> is superior to other <term> kernel
tech,14-4-P03-1005,ak Kernel </term> is superior to other <term> kernel functions </term> and <term> baseline methods </term> .
tech,17-4-P03-1005,ak other <term> kernel functions </term> and <term> baseline methods </term> . Previous research has demonstrated
skrij podrobnosti