manning schuetze statisticalnlp phần 6 ppsx

... tag AT BEZ IN NN VI3 PERIOD 0 0 0 4 863 6 0 19 1973 0 4 26 187 0 38 43322 0 1325 17314 0 185 1 067 3720 42470 11773 61 4 21392 60 72 42 4758 14 76 129 1522 80 16 75 465 6 1329 954 0 First tag AT BEZ IN NN VB PERIOD Table ... VBG DT 99.4 .3 .3 IN .4 97.5 1.5 .5 JJ .l 93.9 1.8 .9 .l .4 NN 2.2 95.5 .2 .4 RB .2 2.4 2.2 .6 93.2 1.2 RP 24.7 1.1 12 .6 61.5 VB .3 1.4 96. 0 VBG 2.5 4.4 93.0...

Ngày tải lên: 14/08/2014, 08:22

... 1992c: 4 76) , so no direct improvement was achieved by clustering. However, a linear interpolation (see section 6. 3.1) between the word-based and the cluster- based model had a perplexity of 2 36, which ... (19 96) , Wang and Waibel (1997), and NieBen et al. (1998). Alshawi et al. (1997), Wang and Waibel (1998), and Wu and Wong 14.1 Hierarchical Clustering 505 0 1 2 3 4 5 6...

Ngày tải lên: 14/08/2014, 08:22

... by simulated annealing. Science 220 :67 1 -68 0. Klavans, Judith, and Min-Yen Kan. 1998. Role of verbs in document analysis. In ACL 36/ COLING 17, pp. 68 0 -68 6. Klavans, Judith L., and Evelyne Tzoukermann. ... marginal grammaticality, 9 A* search, 442 Abney (1991), 3 76, 414,451 Abney (1996a), 375, 3 76 Abney (1996b), 34 Abney (19971,457 absolute discounting, 2 16 accuracy, 269...

Ngày tải lên: 14/08/2014, 08:22

... SBAR NP NP PP 12 Probabilistic Parsing % as Subj % as Obj 13.7% 2.1% 3.5% 0.9% 5 .6% 4 .6% 1.4% 2.8% 0.5% 2 .6% 5 .6% 14.1% Table 12.3 Selected common expansions of NP as Subject vs. Object, ordered by ... theses (Horning 1 969 ; Ellis 1 969 ; Hutchins 1970). Tree structures in probability theory are normally referred to as BRANCHING processes, and are discussed in such work as (Harri...

Ngày tải lên: 21/08/2013, 10:59

... which words and utterances are used. List of Tables 15.3 15.4 15.5 15 .6 15.7 15.8 15.9 16. 1 16. 2 16. 3 16. 4 16. 5 16. 6 16. 7 16. 8 16. 9 Three quantities that are commonly used in term weighting in information ... rankings. 535 349 352 359 363 363 374 375 383 384 394 413 418 420 420 4 36 455 470 500 501 503 518 521 List of Tables xvii 6. 8 6. 9 6. 10 6. 11 7.1 7.2 7.3...

Ngày tải lên: 14/08/2014, 08:22

... lo7 2 8 244.42 4 16 1.34 x 10s 3 7 36. 21 6 14 2307. 06 4 6 7.54 8 12 87.89 5 5 2. 16 10 10 6. 89 6 4 0.84 12 8 1.09 7 3 0.45 14 6 0.35 8 2 0. 36 16 4 0.25 9 1 0.37 18 2 0.48 10 0 0 .68 20 0 3.74 Table ... that: H(X) = H(X) - H(XIX) = 1(X; X) 6. Mutual information is conventionally written with a semi-colon separating the two ar- guments. We are unsure why. 68 2 Ma...

Ngày tải lên: 14/08/2014, 08:22

... different tag sets. Tag set Basic size Total tags Brown 87 179 Penn 45 CLAWS1 132 CLAWS2 166 CLAWS c5 62 London-Lund 197 Table 4.4 Sizes of various tag sets. C5TAGSET Corpus (CLAWS1 through CLAWS5; ... written Amer- ican English from 1 961 . It was compiled and documented by W. Nelson Francis and Henry Kucera (Francis and Kucera 1 964 ; Kucera and Fran- cis 1 967 ; Francis and Kucera 198...

Ngày tải lên: 14/08/2014, 08:22

... 209 It e .s n jt ,f e, :h 1s c3 1s id of ge ce ‘ul System 1 System 2 scores 71, 61 , 55, 60 , 68 , 49, 42, 55, 75, 45, 54, 51 42, 72, 76, 55, 64 55, 36, 58, 55, 67 total 60 9 5 26 n 11 11 mean Xi 55.4 47.8 Sf = C(Xi,j - Xi)2 1,375.4 1,228.g df 10 10 Pooled ... meaningful 214 6 Statistical Inference: n-gram Models over Sparse Data Bigrams r N, r 1 138741 28 2 25413 29 3 10531...

Ngày tải lên: 14/08/2014, 08:22

... sauce .732 pepper .728 salt .7 26 cup .7 26 fallen fell .932 decline .931 rise .930 drop .929 engineered genetically .758 drugs .68 8 research .68 7 drug .68 5 I Alfred named .814 Robert .809 ... .68 5 I Alfred named .814 Robert .809 William .808 W .808 simple something . 964 things . 963 You . 963 always . 962 Table 8.8 The cosine as a measure of semantic similarity. Fo...

Ngày tải lên: 14/08/2014, 08:22

... 784.0 544.0 400.0 3 56. 0 302.0 255.0 1248.0 est. 71092.0 1904.7 1 462 .5 1122.9 862 .2 66 2.1 508.3 390.3 299.7 230.1 students act. 74343.0 2523.0 761 .0 413.0 265 .0 178.0 143.0’ 112.0 96. 0 462 .0 est. 74343.0 ... formula: log O(d) = log p ’; y;;;, 5 86 16 Text Categorization “e arnings” “e arnings ” correct? assigned? YES NO YES- I 1 1024 69 NO 1 63 2143 Table 16. 5 Cont...

Ngày tải lên: 14/08/2014, 08:22