Fig 1.
System overview.
Table 1.
Word and distance feature vector.
Fig 2.
Work flow for feature extraction in both feature based kernel and DSTK.
Fig 3.
Distributed Smoothed Tree (DST) A) Lexicalized parse tree for DST B) Subtrees of tree in lexicalized parse tree.
Table 2.
List of corpora used for evaluation.
Fig 4.
ROC plotting of our three different kernels (feature based, DSTK, Composite) in five corpora a) AIMed b) BioInfer c) HPRD50 d) IEPA e) LLL.
Table 3.
Experimental results on three kernel feature based (Kfea), DSTK (KDSTK) and composite (Kckl).
Table 4.
Comparison of our method with (Li et al [30]) in AIMed Corpus.
Table 5.
Comparison of our method with other kernel based methods.
Table 6.
Comparison of our method with other non-kernel methods.
Table 7.
Complex sentences extracted while annotating PPI.