Improving the identification of non-anaphoric

1
Improving the identification of non-anaphoric
it using support vector machines

• Jose C. Clemente1, Kentaro Torisawa2, Kenji
  Satou1
• 1 Graduate School of Knowledge Science. Genetic
  Knowledge Systems Lab.
• 2 Graduate School of Information Science.
  Knowledge Engineering Lab.
• Japan Advanced Institute of Science and
  Technology. 1-1 Asahidai, Tatsunokuchi. Ishikawa,
  Japan

Fig. 3. Accuracy and Mcc with pattern-based, MBL
and SVM approaches
Fig. 1. (A) Anaphoric it identification and
resolution phases (B) non-anaphoric it no
resolution phase.
Fig. 4. Attribute relevance (A) with MBL (B) with
SVM
Jose C. Clemente
Kenji Satou
Kentaro Torisawa
Fig. 2. Identification of non-anaphoric it using
SVM.