File size: 37,935 Bytes
6fa4bc9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
{
    "paper_id": "C02-1019",
    "header": {
        "generated_with": "S2ORC 1.0.0",
        "date_generated": "2023-01-19T12:20:39.855691Z"
    },
    "title": "The Grammatical Function Analysis between Korean Adnoun Clause and Noun Phrase by Using Support Vector Machines",
    "authors": [
        {
            "first": "Songwook",
            "middle": [],
            "last": "Lee",
            "suffix": "",
            "affiliation": {},
            "email": ""
        },
        {
            "first": "Tae-Yeoub",
            "middle": [],
            "last": "Jang",
            "suffix": "",
            "affiliation": {},
            "email": ""
        },
        {
            "first": "Jungyun",
            "middle": [],
            "last": "Seo",
            "suffix": "",
            "affiliation": {},
            "email": ""
        }
    ],
    "year": "",
    "venue": null,
    "identifiers": {},
    "abstract": "This study aims to improve the performance of identifying grammatical functions between an adnoun clause and a noun phrase in Korean. The key task is to determine the relation between the two constituents in terms of such functional categories as subject, object, adverbial, and appositive. The problem is mainly caused by the fact that functional morphemes, which are considered to be crucial for identifying the relation, are frequently omitted in the noun phrases. To tackle this problem, we propose to employ the Support Vector Machines(SVM) in determining the grammatical functions. Through an experiment with a tagged corpus for training SVMs, the proposed model is found to be useful.",
    "pdf_parse": {
        "paper_id": "C02-1019",
        "_pdf_hash": "",
        "abstract": [
            {
                "text": "This study aims to improve the performance of identifying grammatical functions between an adnoun clause and a noun phrase in Korean. The key task is to determine the relation between the two constituents in terms of such functional categories as subject, object, adverbial, and appositive. The problem is mainly caused by the fact that functional morphemes, which are considered to be crucial for identifying the relation, are frequently omitted in the noun phrases. To tackle this problem, we propose to employ the Support Vector Machines(SVM) in determining the grammatical functions. Through an experiment with a tagged corpus for training SVMs, the proposed model is found to be useful.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Abstract",
                "sec_num": null
            }
        ],
        "body_text": [
            {
                "text": "Many structural ambiguities in Korean sentences are one of the major problems in Korean syntactic analyses. Most of those ambiguities can be classified into either of two categories known as \"noun phrase (NP) attachment problem\" and \"verb phrase (VP) attachment problem\". The NP attachment problem refers to finding the VP which is the head of an NP. On the other hand, the VP attachment problem refers to finding the VP which is the head of a VP.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "In resolving the NP attachment problem, functional morphemes play an important role as they are the crucial elements in characterizing the grammatical function between an NP and its related VP. However, the problem is that there are many NPs that do not have such functional morphemes explicitly attached to each of them. This omission makes it difficult to identify the relation between constituents and subsequently to solve the NP attachment problem. Moreover, most Korean sentences are complex sentences, which also makes the problem more complicated.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "In this research, we make an attempt to solve this problem. The focus is on the analysis of the grammatical function between an NP and an embedded adnoun clause with a functional morpheme omitted.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "We adopt Support Vector Machines(SVM) as the device by which a given adnoun clause is analyzed as one of three relative functions (subject, object, or adverbial) or an appositive. Later in this paper (section 3), a brief description of SVM will be given.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "Adnoun clauses are very frequent in Korean sentences. In a corpus, for example, they appear as often as 18,264 times in 11,932 sentences (see section 4, for details). It means that effective analyses of adnoun clauses will directly lead to improved performance of lexical, morphological and syntactic processing by machine. In order to indicate the difficulties of the adnoun clause analysis, we need to have some basic knowledge on the structure of Korean adnoun clause formation. Thus, we will briefly illustrate the types of Korean adnoun clauses. Then, what makes the analysis tricky will be made clear.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Korean Adnoun Clauses and their analysis problems",
                "sec_num": "2"
            },
            {
                "text": "There are two types of adnoun clauses in Korean : relative adnoun clause and appositive adnoun clause. The former is a more general form of adnoun clause and its formation can be exemplified as follows :",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Two types of adnoun clauses",
                "sec_num": "2.1"
            },
            {
                "text": "1.a Igeos-eun(this) geu-ga(he) sseu-n(wrote) chaeg-ida (book-is) .",
                "cite_spans": [
                    {
                        "start": 55,
                        "end": 64,
                        "text": "(book-is)",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Two types of adnoun clauses",
                "sec_num": "2.1"
            },
            {
                "text": "(This is the book which he wrote.)",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Two types of adnoun clauses",
                "sec_num": "2.1"
            },
            {
                "text": "1.b Igeos-eun(this) chaeg-ida(book-is) .",
                "cite_spans": [
                    {
                        "start": 4,
                        "end": 38,
                        "text": "Igeos-eun(this) chaeg-ida(book-is)",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Two types of adnoun clauses",
                "sec_num": "2.1"
            },
            {
                "text": "(This is a book.)",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Two types of adnoun clauses",
                "sec_num": "2.1"
            },
            {
                "text": "1.c Geu-ga(he) chaeg-eul(book) sseoss-da(wrote).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Two types of adnoun clauses",
                "sec_num": "2.1"
            },
            {
                "text": "(He worte the book.)",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Two types of adnoun clauses",
                "sec_num": "2.1"
            },
            {
                "text": "1.a is a complex sentence composed of two simple sentences 1.b and 1.c in terms of adnoun clause formation. The functional morpheme 'eul', which represents the object relation between 'chaeg' and 'sseoss-da' in 1.c, does not appear in 1.a but 'chaeg' is the functional object of 'sseu-n' in 1.a. This adnoun clause is called a relative adnoun clause whose complement moves to the NP modified by the adnoun clause and the NP modified by a relative adnoun clause is called a head NP. In 1.a 'geu-ga sseun' is a relative adnoun clause and 'chaeg' is its head noun (or NP). Let us consider another example of an adnoun clause.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Two types of adnoun clauses",
                "sec_num": "2.1"
            },
            {
                "text": "Geu-ga(he) jeongjigha-n(be honest) sasil-eun(fact) modeun(every) saram-i(body) an-da(know).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "2.",
                "sec_num": null
            },
            {
                "text": "(Everybody knows the fact that he is honest.)",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "2.",
                "sec_num": null
            },
            {
                "text": "The adnoun clause in 2 is a complete sentence which has all necessary syntactic constituents in itself. This type of adnoun clause is called an appositive adnoun clause. And the head NP modified by the appositive adnoun clause is called a complement noun (Lee, 1986; Chang 1995) . In 2, 'geu-ga jeongjig-han' is an appositive adnoun clause and 'sasil' is a complement noun. Generally, such words as \"iyu(reason), gyeong-u(case), jangmyeon(scene), il(work), cheoji(condition), anghwang(situation), saggeon(happening), naemsae(smell), somun(rumor) and geos(thing)\" are typical examples of the complement noun (Chang, 1995; Lee, 1986 ).",
                "cite_spans": [
                    {
                        "start": 255,
                        "end": 266,
                        "text": "(Lee, 1986;",
                        "ref_id": "BIBREF1"
                    },
                    {
                        "start": 267,
                        "end": 278,
                        "text": "Chang 1995)",
                        "ref_id": "BIBREF0"
                    },
                    {
                        "start": 607,
                        "end": 620,
                        "text": "(Chang, 1995;",
                        "ref_id": "BIBREF0"
                    },
                    {
                        "start": 621,
                        "end": 630,
                        "text": "Lee, 1986",
                        "ref_id": "BIBREF1"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "2.",
                "sec_num": null
            },
            {
                "text": "The first problem we are faced with when analyzing grammatical functions of Korean adnoun clauses is obviously the disappearance of the functional morphemes which carry important information, as shown in the previous subsection (2.1).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The problems",
                "sec_num": "2.2"
            },
            {
                "text": "Apart from the morpheme-ommission problem, there is another reason for the difficulty. As it is directly related to a language particular syntactic characteristic of Korean, we need first to understand a unique procedure of Korean relativization. Unlike English, in which relative pronouns (e.g., who, whom, whose, which and that) are used for relativization and they themselves bear crucial information for identifying grammatical function of the head noun in relative clauses (see example 1.a, in section 1), there is no such relative pronouns in Korean. Instead, an adnominal verb ending is attached to the verb stem and plays a grammatical role of modifying its head noun. However, the problem is that these verb ending morphemes do not provide any information about the grammatical function associated with the relevant head noun.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The problems",
                "sec_num": "2.2"
            },
            {
                "text": "Take 3.a-c for examples. The word order gives little information because Korean is a partly free word-order language and some complements of a verb may be frequently omitted. For example, in sentence 4, the verb of relative clause 'sigdang-eseo meog-eun(who ate in the restaurant or which one ate in the restaurant)' have two omitted complements which are subject and object. So 'bab' can be identified as either of subject or object in the relative clause.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The problems",
                "sec_num": "2.2"
            },
            {
                "text": "4. Sigdang-eseo(restaurant) meog-eun(ate) bab-eul(rice) na-neun(I) boass-da(saw). (I saw the rice which (one) ate in a restaurant.)",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The problems",
                "sec_num": "2.2"
            },
            {
                "text": "Korean appositive adnoun clauses have the same syntactic structure of relative adnoun clauses as in example 2 in section 2.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The problems",
                "sec_num": "2.2"
            },
            {
                "text": "Yoon et al. (1997) classified adnoun clauses into relative adnoun clauses and appositive adnoun clauses based on a complement noun dictionary which was manually constructed, and then tries to find the grammatical function of a relative noun using lexical co-occurrence information. But as shown in example 5, a complement noun can be used as a relative noun, so Yoon et al. (1997)'s method using the dictionary has some limits.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The problems",
                "sec_num": "2.2"
            },
            {
                "text": "Geu-ga(he) balgyeonha-n(discover) sasil-eul(truth) mal-haess-da(talk). (He talked about the truth which he discovered.)",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "5.",
                "sec_num": null
            },
            {
                "text": "Li et al. (1998) described a method using conceptual co-occurrence patterns and syntactic role distribution of relative nouns. Linguistic information is extracted from corpus and thesaurus. However, he did not take into account appositive adnoun clauses but only considered relative adnoun clauses.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "5.",
                "sec_num": null
            },
            {
                "text": "Lee et al. 2001classified adnoun clauses into appositive clauses and one of relative clauses. He proposed a stochastic method based on a maximum likelihood estimation and adopted the backed-off model in estimating the probability P(r|v,e,n) to handle sparse data problem (the symbols r, v, e and n represent the grammatical relation, the verb of the adnoun clause, the adnominal verb ending, and the head noun modified by an adnoun clause, respectively). The backed-off model handles unknown words effectively but it may not be used with all the backed-off stages in real field problems where higher accuracy is needed.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "5.",
                "sec_num": null
            },
            {
                "text": "The technique of Support Vector Machines(SVM) is a learning approach for solving two-class pattern recognition problems introduced by Vapnik (1995) . It is based on the Structural Risk Minimization principle for which error-bound analysis has been theoretically motivated (Vapnik, 1995) . The problem is to find a decision surface that separates the data points in two classes optimally. A decision surface by SVM for linearly separable space is a hyperplane H : y = w\u22c5x -b = 0 and two hyperplanes parallel to it and with equal distances to it, H 1 : y = w\u22c5x -b = +1, H 2 : y = w\u22c5x -b = -1, with the condition that there are no data points between H 1 and H 2 , and the distance between H 1 and H 2 is maximized.",
                "cite_spans": [
                    {
                        "start": 134,
                        "end": 147,
                        "text": "Vapnik (1995)",
                        "ref_id": "BIBREF3"
                    },
                    {
                        "start": 272,
                        "end": 286,
                        "text": "(Vapnik, 1995)",
                        "ref_id": "BIBREF3"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Support Vector Machines",
                "sec_num": "3"
            },
            {
                "text": "We want to maximize the distance between H 1 and H 2 . So there will be some positive examples on H 1 and some negative examples on H 2 . These examples are called support vectors because they only participate in the definition of the separating hyperplane, and other examples can be removed and/or moved around as long as they do not cross the planes H 1 and H 2 . In order to maximize the distance, we should minimize ||w|| with the condition that there are no data points between H 1 and H 2 ,",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Support Vector Machines",
                "sec_num": "3"
            },
            {
                "text": "w\u22c5x -b \u2265 +1 for y i = +1, w\u22c5x -b \u2264 -1 for y i = -1.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Support Vector Machines",
                "sec_num": "3"
            },
            {
                "text": "The SVM problem is to find such w and b that satisfy the above constraints. It can be solved using quadratic programming techniques (Vapnik, 1995) . The algorithms for solving linearly separable cases can be extended so that they can solve linearly non-separable cases as well by either introducing soft margin hyperplanes, or by mapping the original data vectors to a higher dimensional space where the new features contain interaction terms of the original features, and the data points in the new space become linearly separable (Vapnik, 1995) . We use SVM light 1 system for our experiment (Joachimes, 1998) .",
                "cite_spans": [
                    {
                        "start": 132,
                        "end": 146,
                        "text": "(Vapnik, 1995)",
                        "ref_id": "BIBREF3"
                    },
                    {
                        "start": 532,
                        "end": 546,
                        "text": "(Vapnik, 1995)",
                        "ref_id": "BIBREF3"
                    },
                    {
                        "start": 594,
                        "end": 611,
                        "text": "(Joachimes, 1998)",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Support Vector Machines",
                "sec_num": "3"
            },
            {
                "text": "SVM performance is governed by the features. We use the verb of each adnoun clause, the adnominal verb ending and the head noun of the noun phrase. To reflect context of sentence, we use the previous noun phrase, which is located right before the verb, and its functional morpheme. The previous noun phrase is the surface level word list not the previous argument for the verb in adnoun clause. Part of speech(POS) tags of all lexical item are also used as feature. For example, in sentence 'Igeos-eun geu-ga sseu-n chaeg-ida.', 'geu' is a previos noun pharse feature, 'ga' is its functional morpheme feature, 'sseu' is a verb feature, 'n' is a verb ending feature, 'chaeg' is a head noun feature and all POS tags of lexical items are features.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Support Vector Machines",
                "sec_num": "3"
            },
            {
                "text": "Because we found that the kernel of SVM does not strongly affect the performance of our problem through many experiments, we concluded that our problem is linearly separable. Thus we will use the linear kernel only.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Support Vector Machines",
                "sec_num": "3"
            },
            {
                "text": "As the SVMs is a binary class classifier, we construct four classifiers, one for each class. Each classifier constructs a hyperplane between one class and other classes. We select the classifier which has the maximal distance from the margin for each test data point.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Support Vector Machines",
                "sec_num": "3"
            },
            {
                "text": "We use the tree tagged corpus of Korean Information Base which is annotated as a form of phrase structured tree (Lee, 1996) . It consists of 11,932 sentences, which corresponds to 145,630 eojeols. Eojeol is a syntactic unit composed of one lexical morpheme with multiple functional morphemes optionally attached to it. We extract the verb of an adnoun clause and the noun phrase which is modified by the adnoun clause. We regard an eojeol consisting of a main verb and auxiliary-verbs as a single main-verb eojeol. In case of a complex verb, we only take into account the first part of it. Every verb which has adnominal morphemes and the head word of a noun phrase which is modified by adnoun clause, were extracted. Because Korean is head-fiinal 1",
                "cite_spans": [
                    {
                        "start": 112,
                        "end": 123,
                        "text": "(Lee, 1996)",
                        "ref_id": "BIBREF1"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Experimental Results",
                "sec_num": "4"
            },
            {
                "text": "The SVMlight system is available at http://ais.gmd.de/~thorsten/svm_light/. language, we regard the last noun of a noun phrase as the head word of the noun phrase.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Experimental Results",
                "sec_num": "4"
            },
            {
                "text": "The total number of extracted pairs of verb and noun is 18,264. The grammatical function of each pair is manually tagged. To experiment, the data was subdivided into a learning data set from 10,739 sentences and a test data set from 1,193 sentences. We use 16,413 training data points and 1,851 test data points in all experiments. Table 1 shows an accuracy at each of the grammatical categories between an adnoun clause and a noun phrase with SVMs, compared with the backed-off method which is proposed by (Lee, 2001 ). It should be noted that SVM outperforms Backed-off model in Table 1 . By using context information we acquire an improvement of overall 2.1%. Table 2 represents the accuracies of the proposed model compared with the Li's model. The category 'appositive' is not taken into account for fair comparison. It should be noted that Li et al. (1998) 's results are drawn from most frequent 100 verbs while ours, from 4,684 verbs all of which are in the training corpus. It is shown that our proposed model shows the better overall result in determining the grammatical function between an adnoun clause and its modifying head noun. Most errors are caued by lack of lexical information. Actually, lexical information in 19% of the test data has not occurred in the training data. The other errors are caused by the characteristics that some verbs in adnoun clauses can have dual subjects which we did not consider in the problem. Take 6 for an example.",
                "cite_spans": [
                    {
                        "start": 507,
                        "end": 517,
                        "text": "(Lee, 2001",
                        "ref_id": "BIBREF1"
                    },
                    {
                        "start": 846,
                        "end": 862,
                        "text": "Li et al. (1998)",
                        "ref_id": "BIBREF2"
                    }
                ],
                "ref_spans": [
                    {
                        "start": 332,
                        "end": 339,
                        "text": "Table 1",
                        "ref_id": "TABREF1"
                    },
                    {
                        "start": 581,
                        "end": 588,
                        "text": "Table 1",
                        "ref_id": "TABREF1"
                    },
                    {
                        "start": 663,
                        "end": 670,
                        "text": "Table 2",
                        "ref_id": "TABREF2"
                    }
                ],
                "eq_spans": [],
                "section": "Experimental Results",
                "sec_num": "4"
            },
            {
                "text": "(Cheolsu who has big eyes)",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Nun-i(eyes) keu-n(be big) Cheolsu",
                "sec_num": "6."
            },
            {
                "text": "In example 6, the context NP is 'nun' and its functional word is 'i' which may represent that it is subject of 'keu-da', thus system may wrongly determine that 'Cheolsu' is not a subject of 'keu-da' because the subject of 'keu-da' has been made with 'nun'. However, both 'Cheolsu' and 'nun' are the subjects of 'keu-da'.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Nun-i(eyes) keu-n(be big) Cheolsu",
                "sec_num": "6."
            },
            {
                "text": "Adnoun clause is a typical complex sentence structure of Korean. There are various types of grammatical relations between an adnoun clause and its relevant noun phrase. Unlike in between general content words and modifying clauses where their grammatical relations can be easily extrated in terms of various grammatical characteristics by the functional morphemes, the functional morphemes are omitted in a noun phrase when it is modified by an adnoun clause. This omission makes it difficult to characterize their grammatical relation. In this paper, we used SVM to take care of this problem and analyze the relation between noun phrase and adnoun clause. We reflected context information by using the previous word of the verb in adnoun clauses as feature. Context information helped the grammatical function analysis between adnoun clause and the head noun. The SVM can also handle the sparse data problem as the backed-off model does. We acquired overall accuracy of 90.8%, which is obviously an improvement from the previous works.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Conclusion and Future works",
                "sec_num": "5"
            },
            {
                "text": "In the future, we plan to compare with other machine learning methods and to enhance our system by using a publicly available Korean thesaurus to increases general accuracy. More data needs to be collected for further performance improvement. We will also work on utilizing the proposed model in some partial parsing problem.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Conclusion and Future works",
                "sec_num": "5"
            }
        ],
        "back_matter": [],
        "bib_entries": {
            "BIBREF0": {
                "ref_id": "b0",
                "title": "Syntactic Analysis for Korean Sentences Using Lexical Association Based on Co-occurrence Relation",
                "authors": [
                    {
                        "first": "Suk-Jin",
                        "middle": [],
                        "last": "Chang",
                        "suffix": ""
                    }
                ],
                "year": 1995,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Chang, Suk-Jin, 1995. Information-based Korean Grammar, Hanshin Publishing Co. Yoon, J., 1997. Syntactic Analysis for Korean Sentences Using Lexical Association Based on Co-occurrence Relation, Ph.D. Dissertation, Yonsei University.",
                "links": null
            },
            "BIBREF1": {
                "ref_id": "b1",
                "title": "Estimation of Probabilities from Sparse Data for the Language Model Component of a Speech Recogniser",
                "authors": [
                    {
                        "first": "S",
                        "middle": [],
                        "last": "Katz",
                        "suffix": ""
                    },
                    {
                        "first": "",
                        "middle": [],
                        "last": "Hagyeonsa",
                        "suffix": ""
                    },
                    {
                        "first": "Kong",
                        "middle": [],
                        "last": "Lee",
                        "suffix": ""
                    },
                    {
                        "first": "Jae-Hoon",
                        "middle": [],
                        "last": "Joo",
                        "suffix": ""
                    },
                    {
                        "first": "Key-Sun",
                        "middle": [],
                        "last": "Kim",
                        "suffix": ""
                    },
                    {
                        "first": "Gil",
                        "middle": [
                            "Chang"
                        ],
                        "last": "Choi",
                        "suffix": ""
                    },
                    {
                        "first": "",
                        "middle": [],
                        "last": "Kim",
                        "suffix": ""
                    }
                ],
                "year": 1986,
                "venue": "Proceedings of the Sixth Natural Language Processing Pacific Rim Symposium",
                "volume": "35",
                "issue": "",
                "pages": "709--713",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Katz, S., 1987. Estimation of Probabilities from Sparse Data for the Language Model Component of a Speech Recogniser. IEEE Transactions on Acoustics, Speech, and Signal processing, Vol. ASSP-35, No. 3. Lee, Ik-Sop, Hong-Pin Im, 1986, Korean Grammar Theory, Hagyeonsa. Lee, Kong Joo, Jae-Hoon Kim, Key-Sun Choi, and Gil Chang Kim. 1996, Korean syntactic tagset for building a tree annotated corpus. Korean Journal of Cognitive Science, 7(4):7-24. Lee, Songwook, Tae-Yeoub Jang, Jungyun Seo. 2001, The Grammatical Function Analysis between Adnoun Clause and Noun Phrase in Korean, In Proceedings of the Sixth Natural Language Processing Pacific Rim Symposium, pp709-713.",
                "links": null
            },
            "BIBREF2": {
                "ref_id": "b2",
                "title": "Identifying Syntactic Role of Antecedent in Korean Relative Clause Using Corpus and Thesaurus Information",
                "authors": [
                    {
                        "first": "Hui-Feng",
                        "middle": [],
                        "last": "Li",
                        "suffix": ""
                    },
                    {
                        "first": "Jong-Hyeok",
                        "middle": [],
                        "last": "Lee",
                        "suffix": ""
                    },
                    {
                        "first": "Geunbae",
                        "middle": [],
                        "last": "Lee",
                        "suffix": ""
                    }
                ],
                "year": 1998,
                "venue": "Proceeding of COLING-ACL",
                "volume": "",
                "issue": "",
                "pages": "756--762",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Li, Hui-Feng, Jong-Hyeok Lee, Geunbae Lee, 1998. Identifying Syntactic Role of Antecedent in Korean Relative Clause Using Corpus and Thesaurus Information. In Proceeding of COLING-ACL, pp.756-762.",
                "links": null
            },
            "BIBREF3": {
                "ref_id": "b3",
                "title": "Text Categorization with Support Vector Machines: Learning with Many Relevant Features",
                "authors": [
                    {
                        "first": "Vladimir",
                        "middle": [
                            "N"
                        ],
                        "last": "Vapnik",
                        "suffix": ""
                    }
                ],
                "year": 1995,
                "venue": "The Nature of Statistical Learning Theory",
                "volume": "",
                "issue": "",
                "pages": "137--142",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Vapnik, Vladimir N. 1995, The Nature of Statistical Learning Theory. Springer, New York. Joachims, Thorsten. 1998, Text Categorization with Support Vector Machines: Learning with Many Relevant Features. In European Conference on Machine Learning, pp. 137-142.",
                "links": null
            }
        },
        "ref_entries": {
            "TABREF1": {
                "content": "<table><tr><td colspan=\"5\">the acuracy of SVM and Backed-off</td></tr><tr><td colspan=\"5\">model at each of the grammatical categories</td></tr><tr><td/><td>subj</td><td>obj</td><td colspan=\"2\">adv app total</td></tr><tr><td>SVM</td><td>84.4</td><td colspan=\"3\">62.9 92.0 97.5 88.7</td></tr><tr><td>SVM with</td><td/><td/><td/><td/></tr><tr><td>context</td><td>88.8</td><td colspan=\"3\">75.6 89.6 96.1 90.8</td></tr><tr><td>feature</td><td/><td/><td/><td/></tr><tr><td>Backed-off</td><td>86.2</td><td colspan=\"3\">42.0 62.0 91.7 83.5</td></tr><tr><td>proportion in</td><td/><td/><td/><td/></tr><tr><td>the training</td><td>52.8</td><td>4.5</td><td>6.7</td><td>36.0 100</td></tr><tr><td>data(%)</td><td/><td/><td/><td/></tr></table>",
                "type_str": "table",
                "html": null,
                "num": null,
                "text": ""
            },
            "TABREF2": {
                "content": "<table><tr><td colspan=\"5\">the accuracy of SVM without</td></tr><tr><td colspan=\"3\">considering appositive clauses</td><td/></tr><tr><td/><td>subj</td><td>obj</td><td>adv</td><td>total</td></tr><tr><td>SVM with</td><td/><td/><td/></tr><tr><td>context</td><td colspan=\"4\">94.1 87.8 85.7 93.3</td></tr><tr><td>feature</td><td/><td/><td/></tr><tr><td>Li et al. (1998)</td><td>90</td><td>92</td><td colspan=\"2\">89.2 90.4</td></tr></table>",
                "type_str": "table",
                "html": null,
                "num": null,
                "text": ""
            }
        }
    }
}