3. Precision
5:
(3) ORG PER LOC NOTNE
ORG 261 22 20 76
F PER 9 103 4 24
2 ∗ P recision ∗ Recall LOC 14 5 41 5
(4)
P recision + Recall NOTNE 216 54 11 633
6:
5 ORG PER LOC NOTNE
ORG 220 12 11 136
5.1 PER 12 97 1 30
LOC 11 1 43 10
, 3
. ,ORG,PER,LOC ,Recall NOTNE 105 18 12 779
Precision .
Recall , Precision 7:
. , ORG PER LOC NOTNE
, F 70 ORG 235 7 8 129
. PER 5 106 1 28
LOC 13 3 43 6
3: NOTNE 89 12 4 809
Acc Rec Pre F
69.29 69.35 53.29 60.27
76.03 61.64 66.30 63.89 6.1
79.64 65.75 73.00 69.19
6.1.1
,
,
. 8 , 10
5.2 100 . ”
” ” ” , 10 ,
, 4 .
NOTNE
,
,100 . ”
Recall, Precision .
” ,100
,ORG Recall ,
. , ,PER
.
, ORG
.
8:
4: 10 100
PER NOTNE PER
Rec Pre F Rec Pre F Rec Pre F
ORG 68.87 52.20 59.39 58.05 63.22 60.52 62.01 68.71 65.19 LOC NOTNE LOC
PER 73.57 55.98 63.58 69.29 75.78 72.39 75.71 82.81 79.10 NOTNE NOTNE PER
LOC 63.08 53.95 58.16 66.15 64.18 65.15 66.15 76.79 71.07
2 ,
, . Recall
. ,
5.3 ,
. ,
5,6,7 , ,
. .
, ORG
. NOTNE ORG PER
,
. 6.1.2
,
, , ,
. . ,
,
, , .
, ,
.
6 ,
, , .
.
4. 75
11:
70
(C= ,P= ,S= )
65 ORG vs PER PER vs ORG LOC vs ORG NOTNE vs ORG
C: C: S: C:
60 P: S: S: C:
Recall C: C: S: C:
55
Precision C: S: P: C:
50 C: C: S: C:
ORG vs LOC PER vs LOC LOC vs PER NOTNE vs PER
45 C: C: S: C:
C: S: S: C:
40 C: C: P: C:
10 20 30 40 50 60 70 80 90 100 110 120 C: C: P: P:
P: S: S: C:
ORG vs NOTNE PER vs NOTNE LOC vs NOTNE NOTNE vs LOC
C: P: S: C:
C: P: S: C:
2: C: P: C: P:
P: C: P: C:
C: S: S: P:
9 ”arsenal” ” ” ,
,
,
. ,
7
, ,
. .
, .
, Recall
9: . ,
Precision .
arsenal ORG NOTNE
PER NOTNE
LOC LOC
ORG ORG
[1] B.J. Jansen, A. Spink, and T. Saracevic. Real life,
real users, and real needs: a study and analysis of
user queries on the web. Information processing &
management, Vol. 36, No. 2, pp. 207–227, 2000.
6.2 [2] S.M. Beitzel. On understanding and classifying web
queries. PhD thesis, Citeseer, 2006.
,
, [3] R.E. Fan, K.W. Chang, C.J. Hsieh, X.R. Wang, and
, C.J. Lin. LIBLINEAR: A library for large linear
. classification. The Journal of Machine Learning Re-
, ”arsenal” ” search, Vol. 9, pp. 1871–1874, 2008.
” , . [4] , .
, ,” ” LOC JUMAN version6.0.
ORG http://nlp.kuee.kyoto-u.ac.jp/nl-resource/
. juman.html , 2009.
[5] Y. Li, Z. Zheng, and H.K. Dai. KDD CUP-2005 re-
10: port: Facing a great challenge. ACM SIGKDD Ex-
plorations Newsletter, Vol. 7, No. 2, pp. 91–99, 2005.
arsenal ORG ORG [6] D. Shen, R. Pan, J.T. Sun, J.J. Pan, K. Wu, J. Yin,
PER PER and Q. Yang. Q 2 C@ UST: our winning solu-
tion to query classification in KDDCUP 2005. ACM
LOC ORG SIGKDD Explorations Newsletter, Vol. 7, No. 2, pp.
ORG LOC 100–110, 2005.
[7] , .
11 , .
, 5 , Vol. 15, No. 5, p. 10, 2008.
. ,PER vs ORG , ”
” , ” ” [8] , , . ,
, ORG PER . The 23rd An-
. nual Conference of the Japanese Society for Artificial
,ORG NOTNE Intelligence, 2009.
,LOC [9] IREX Homepage. http://nlp.cs.nyu.edu/irex/
,PER index-j.html.
.