Bernard Lamers - PowerPoint PPT Presentation

1 / 13
About This Presentation
Title:

Bernard Lamers

Description:

8 9 Bernard Lamers Schulte im Walde, Sabine and Brew, Chris. 2002. Inducing German Semantic Verb Classes from Purely ... – PowerPoint PPT presentation

Number of Views:84
Avg rating:3.0/5.0
Slides: 14
Provided by: berna173
Category:
Tags: bernard | chris | lamers | sabine

less

Transcript and Presenter's Notes

Title: Bernard Lamers


1
???8?9?
  • Bernard Lamers

2
?????????
  • Schulte im Walde, Sabine and Brew, Chris. 2002.
    Inducing German Semantic Verb Classes from Purely
    Syntactic Subcategorisation Information.
    Proceedings ACL.

3
Inducing German Semantic Verb Classes
  • ??
  • ???????
  • ??????????????????clustering
  • ??????2500??????????????clustering?k-means??????
  • ????????????????????????

4
??????????
  • ?????????????argument???????argument
    nominative(n), dative(d), accusative(a),
    reflexives(r), prepositional phrases(p),
    expletive es(x), non-finite clauses(i) etc.
  • ?????nai????????????38?????????????

5
????????????
  • glauben (??????)

???? ??
ns-dass 0.27945
ns-2 0.27358
np 0.09951
n 0.08811
na 0.08046
ni 0.05015
6
????????
  • ????????????????????????????????
  • 57?????14????????
  • Announcement ankundigen(????),bekanntgeben(????),
    eroffnen(?????),verkunden(????)
  • Manner of motion fahren(????),fliegen(??),rudern(
    ??),segeln(????)

7
Clustering??
  • K-means??????n????????????????k?????????????????c
    lustering??
  • 1?????????????????????
  • 2????????centroid??????????1????
  • ????????????????????????????

8
K-means???starting clusters???
  • Starting clusters???
  • Random
  • Agglomerative hierarchical clustering
  • ?????????????????
  • ???????????? merge??
  • k???????????????
  • Merge??single-linkage, complete-linkage, average
    verb distance, distance between cluster centroids
    and Wards method

9
?????????
  • ??cosine??????????Kullback-Leibler divergence
    (relative entropy)
  • KL-divergence???
  • Information radius d(v1, v2) D(p (pq)/2 )
    D(q (pq)/2)
  • Skew divergence d(v1 , v2) D(p wq
    (1-w)p)
  • ??????????q???zero value??????????

10
Clustering evaluation MI(A, B)
  • ???cluster???cluster purity ABij????ABij???????Bj?
    ????????Ai????????
  • ????clustering????(4)
  • MI(A, B)?????????????? ?????????APP??????

11
Clustering evaluation APP
  • APP adjusted pairwise precision?
  • ???cluster????(5)
  • ????clustering????(6)?
  • MI???0.229-0.493
  • APP???0.017-0.291

12
???clustering
  • ????1?2
  • ???????clustering??6?
  • ?????????????????clustering????????????

13
??
  • ????????????????????????????
  • ????????syntax????????????????????
  • ??????
  • selectional restrictions??????
Write a Comment
User Comments (0)
About PowerShow.com