File size: 62,273 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 | {
"paper_id": "C02-1012",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T12:19:31.936313Z"
},
"title": "Chinese Named Entity Identification Using Class-based Language Model 1",
"authors": [
{
"first": "Jian",
"middle": [],
"last": "Sun",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Beijing University of Posts & Telecommunications",
"location": {
"country": "China"
}
},
"email": ""
},
{
"first": "Jianfeng",
"middle": [],
"last": "Gao",
"suffix": "",
"affiliation": {},
"email": "jfgao@microsoft.com"
},
{
"first": "Lei",
"middle": [],
"last": "Zhang",
"suffix": "",
"affiliation": {},
"email": ""
},
{
"first": "Ming",
"middle": [],
"last": "Zhou",
"suffix": "",
"affiliation": {},
"email": "mingzhou@microsoft.com"
},
{
"first": "Changning",
"middle": [],
"last": "Huang",
"suffix": "",
"affiliation": {},
"email": "cnhuang@microsoft.com"
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "",
"pdf_parse": {
"paper_id": "C02-1012",
"_pdf_hash": "",
"abstract": [],
"body_text": [
{
"text": "1 This work was done while the author was visiting Microsoft Research Asia",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "We consider here the problem of Chinese named entity (NE) identification using statistical language model(LM). In this research, word segmentation and NE identification have been integrated into a unified framework that consists of several class-based language models. We also adopt a hierarchical structure for one of the LMs so that the nested entities in organization names can be identified. The evaluation on a large test set shows consistent improvements. Our experiments further demonstrate the improvement after seamlessly integrating with linguistic heuristic information, cache-based model and NE abbreviation identification.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "$EVWUDFW",
"sec_num": null
},
{
"text": "1(LGHQWLILFDWLRQ is the key technique in many applications such as information extraction, question answering, machine translation and so on. English NE identification has achieved a great success. However, for Chinese, NE identification is very different. There is no space to mark the word boundary and no standard definition of words in Chinese. The Chinese NE identification and word segmentation are interactional in nature.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": ",QWURGXFWLRQ",
"sec_num": null
},
{
"text": "This paper presents a unified approach that integrates these two steps together using a class-based LM, and apply Viterbi search to select the global optimal solution. The class-based LM consists of two sub-models, namely the context model and the entity model. The context model estimates the probability of generating a NE given a certain context, and the entity model estimates the probability of a sequence of Chinese characters given a certain kind of NE. In this study, we are interested in three kinds of Chinese NE that are most commonly used, namely person name (PER), location name (LOC) and organization name (ORG) . We have also adopted a variety of approaches to improving the LM. In addition, a hierarchical structure for organization LM is employed so that the nested PER, LOC in ORG can be identified.",
"cite_spans": [
{
"start": 620,
"end": 625,
"text": "(ORG)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": ",QWURGXFWLRQ",
"sec_num": null
},
{
"text": "The evaluation is conducted on a large test set in which NEs have been manually tagged. The experiment result shows consistent improvements over existing methods. Our experiments further demonstrate the improvement after integrating with linguistic heuristic information, cache-based model and NE abbreviation identification. The precision of PER, LOC, ORG on the test set is 79.86%, 80.88%, 76.63%, respectively; and the recall is 87.29%, 82.46%, 56.54%, respectively.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": ",QWURGXFWLRQ",
"sec_num": null
},
{
"text": "Recently, research on English NE identification has been focused on the machine-learning approaches, including hidden Markov model (HMM), maximum entropy model, decision tree and transformation-based learning, etc. (Bikel et al, 1997; Borthwick et al, 1999; Sekine et al, 1998) . Some systems have been applied to real application.",
"cite_spans": [
{
"start": 215,
"end": 234,
"text": "(Bikel et al, 1997;",
"ref_id": "BIBREF1"
},
{
"start": 235,
"end": 257,
"text": "Borthwick et al, 1999;",
"ref_id": "BIBREF0"
},
{
"start": 258,
"end": 277,
"text": "Sekine et al, 1998)",
"ref_id": "BIBREF7"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "5HODWHG:RUN",
"sec_num": null
},
{
"text": "Research on Chinese NE identification is, however, still at its early stage. Some researches apply methods of English NE identification to Chinese. Yu et al (1997) applied the HMM approach where the NE identification is formulated as a tagging problem using Viterbi algorithm. In general, current approaches to NE identification (e.g. Chen, 1997) usually contain two separate steps: word segmentation and NE identification. The word segmentation error will definitely lead to errors in the NE identification results. Zhang (2001) put forward class-based LM for Chinese NE identification. We further develop this idea with some new features, which leads to a new framework. In this framework, we integrate Chinese word segmentation and NE identification into a unified framework using a class-based language model (LM).",
"cite_spans": [
{
"start": 148,
"end": 163,
"text": "Yu et al (1997)",
"ref_id": "BIBREF8"
},
{
"start": 335,
"end": 346,
"text": "Chen, 1997)",
"ref_id": "BIBREF4"
},
{
"start": 517,
"end": 529,
"text": "Zhang (2001)",
"ref_id": "BIBREF9"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "5HODWHG:RUN",
"sec_num": null
},
{
"text": "The n-gram LM is a stochastic model which predicts the next word given the previous n-1 words by estimating the conditional probability P(w n |w 1 \u2026w n-1 ).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&ODVVEDVHG/0 IRU1( ,GHQWLILFDWLRQ",
"sec_num": null
},
{
"text": "In practice, trigram approximation P(w i |w i-2 w i-1 ) is widely used, assuming that the word w i depends only on two preceding words w i-2 and w i-1 . Brown et al (1992) put forward and discussed n-gram models based on classes of words. In this section, we will describe how to use class-based trigram model for NE identification. Each kind of NE (including PER, LOC and ORG) is defined as a class in the model. In addition, we differentiate the transliterated person name (FN) from the Chinese person name since they have different constitution patterns. The four classes of NE used in our model are shown in Table 1 Given a Chinese character sequence 6=V \u00abV Q , the task of Chinese NE identification is to find the optimal class sequence &=F \u00abF P (P<=Q) that maximizes the probability 3&_6. It can be expressed in the equation 1and we call it class-based model.",
"cite_spans": [
{
"start": 153,
"end": 171,
"text": "Brown et al (1992)",
"ref_id": "BIBREF2"
}
],
"ref_spans": [
{
"start": 612,
"end": 619,
"text": "Table 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "&ODVVEDVHG/0 IRU1( ,GHQWLILFDWLRQ",
"sec_num": null
},
{
"text": "The class-based model consists of two sub-models: the context model 3& and the entity model P (S|C). The context model indicates the probability of generating a NE class given a (previous) context. P(C) is a priori probability, which is computed according to Equation 2:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&ODVVEDVHG/0 IRU1( ,GHQWLILFDWLRQ",
"sec_num": null
},
{
"text": "\u00d5 = - - @ P L L L L F F F 3 & 3 1 1 2 ) | ( ) (",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&ODVVEDVHG/0 IRU1( ,GHQWLILFDWLRQ",
"sec_num": null
},
{
"text": "(2) P(C) can be estimated using a NE labeled corpus. The entity model can be parameterized by Equation 3:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&ODVVEDVHG/0 IRU1( ,GHQWLILFDWLRQ",
"sec_num": null
},
{
"text": "\u00d5 = - - - - @ @ = P M M HQG F VWDUW F P Q VWDUW F HQG F P Q F V V 3 F F V V V V 3 F F V V 3 & 6 3 M M P 1 1 1 1 1 ) | ] ... ([ ) ... | ] ... ]...[ ... ([ ) ... | ... ( ) | ( 1 (3)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&ODVVEDVHG/0 IRU1( ,GHQWLILFDWLRQ",
"sec_num": null
},
{
"text": "The entity model estimates the generative probability of the Chinese character sequence in square bracket pair (i.e. starting from F M VWDUW to F M HQG) given the specific NE class.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&ODVVEDVHG/0 IRU1( ,GHQWLILFDWLRQ",
"sec_num": null
},
{
"text": "For different class, we define the different entity model.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&ODVVEDVHG/0 IRU1( ,GHQWLILFDWLRQ",
"sec_num": null
},
{
"text": "For the class of PER (including PN and FN), the entity model is a FKDUDFWHUEDVHG trigram model as shown in Equation 4.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&ODVVEDVHG/0 IRU1( ,GHQWLILFDWLRQ",
"sec_num": null
},
{
"text": "\u00d5 - - = - - - - = = = HQG F VWDUW F N M N N N M HQG F VWDUW F M M M M 3(5 F V V V 3 3(5 F V V 3 ) , , | ( ) | ] ... ([ 1 2 (4)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&ODVVEDVHG/0 IRU1( ,GHQWLILFDWLRQ",
"sec_num": null
},
{
"text": "where s can be any characters occurred in a person name. For example, the generative probability of character sequence \"\u00fb (Li Dapeng) is much larger than that of \u00eeH (many years) given the PER since \" is a commonly used family name, and \u00fb and are commonly used first names. The probabilities can be estimated with the person name list.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&ODVVEDVHG/0 IRU1( ,GHQWLILFDWLRQ",
"sec_num": null
},
{
"text": "For the class of LOC, the entity model is a ZRUGEDVHG trigram model as shown in Equation (5).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&ODVVEDVHG/0 IRU1( ,GHQWLILFDWLRQ",
"sec_num": null
},
{
"text": ") | ( max arg * 6 & 3 & & = ) | ( ) ( max arg & 6 3 & 3 &= (1) ) | ] ... ([ /2& F V V 3 M HQG F VWDUW F M M = - - @ /2& F Z Z _ Z 3 > PD[ /2& F _ Z Z 3 PD[ O N M N N N : M O : \u00d5 = - - = = = \u00bb (5)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&ODVVEDVHG/0 IRU1( ,GHQWLILFDWLRQ",
"sec_num": null
},
{
"text": "where W = w 1 \u2026w l is possible segmentation result of character sequence",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&ODVVEDVHG/0 IRU1( ,GHQWLILFDWLRQ",
"sec_num": null
},
{
"text": "HQG F VWDUW F M M V V - - ... .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&ODVVEDVHG/0 IRU1( ,GHQWLILFDWLRQ",
"sec_num": null
},
{
"text": "For the class of ORG, the construction is much more complicated because an ORG often contain PER and/or LOC. For example, the ORG \" \u00d1 \u00d1 \u00fe N ( \u00cc \" (Air China Corporation) contains the LOC \"\u00d1\" (China). It is beneficial to such applications as question answering, information extraction and so on if nested NE can be identified as well . In order to identify the nested PER, LOC in ORG 2 , we adopted class-based LMs for ORG further, in which there are three sub models, one is the class generative model, and the others are entity model: person name model and location name model in ORG. Therefore, the entity model of ORG is shown in Equation 6which is almost same as Equation (1).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&ODVVEDVHG/0 IRU1( ,GHQWLILFDWLRQ",
"sec_num": null
},
{
"text": ") | ] ... ([ 25* F V V 3 M HQG F VWDUW F M M = - - \u00fa \u00fa \u00fa \u00fa \u00fa \u00fb \u00f9 \u00ea \u00ea \u00ea \u00ea \u00ea \u00eb \u00e9 = = @ \u00fa \u00fa \u00fb \u00f9 \u00ea \u00ea \u00eb \u00e9 = = = = @ \u00d5 \u00d5 = - - = - - - - - - N L M L HQG F VWDUW F N L M L L L & M N HQG F VWDUW F M N & M HQG F VWDUW F M & 25* F F V V 3 25* F F F _ 3F 25* F F F V V 3 25* F F F 3 25* F & V V 3 F & 3 L L M M M M 1 ' ' ' 1 1 ' ' ) , ' | ] ... ([ max ) , ' ... ' | ] ... ([ ) | ' ... ' ( max )] , ' | ] ... ([ ) | ' ( [ max (6) where ' ... ' 1 ' N F F & =",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&ODVVEDVHG/0 IRU1( ,GHQWLILFDWLRQ",
"sec_num": null
},
{
"text": "is the sequence of class corresponding to the Chinese character sequence.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&ODVVEDVHG/0 IRU1( ,GHQWLILFDWLRQ",
"sec_num": null
},
{
"text": "In",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&ODVVEDVHG/0 IRU1( ,GHQWLILFDWLRQ",
"sec_num": null
},
{
"text": "EQUATION",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [
{
"start": 0,
"end": 8,
"text": "EQUATION",
"ref_id": "EQREF",
"raw_str": "addition, if M F is a normal word, 1 ) | ] ... ([ = - - M HQG F VWDUW F F V V 3 M M .",
"eq_num": "(7)"
}
],
"section": "&ODVVEDVHG/0 IRU1( ,GHQWLILFDWLRQ",
"sec_num": null
},
{
"text": "Based on the context model and entity models, we can compute the probability 3&_6",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&ODVVEDVHG/0 IRU1( ,GHQWLILFDWLRQ",
"sec_num": null
},
{
"text": "2 For simplification, only nested person, location names are identified in organization. The nested person in location is not identified because of low frequency and can get the optimal class sequence The Chinese PER and transliterated PER share the same context class model when computing the probability.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&ODVVEDVHG/0 IRU1( ,GHQWLILFDWLRQ",
"sec_num": null
},
{
"text": "As discussed in 3.1.1, there are two kinds of probabilities to be estimated: P(C) and P(S|C) . Both probabilities are estimated using Maximum Likelihood Estimation (MLE) with the annotated training corpus.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "0RGHOV(VWLPDWLRQ",
"sec_num": null
},
{
"text": "The parser NLPWin 3 was used to tag the training corpus. As a result, the corpus was annotated with NE marks. Four lists were extracted from the annotated corpus and each list corresponds one NE class. The context model 3& was trained with the annotated corpus and the four entity models were trained with corresponding NE lists. The Figure 1 shows the training process. (Begin of sentence (BOS) and end of sentence (EOS) is added) Given a sequence of Chinese characters, the decoding process consists of the following three steps:",
"cite_spans": [],
"ref_spans": [
{
"start": 334,
"end": 342,
"text": "Figure 1",
"ref_id": "FIGREF0"
}
],
"eq_spans": [],
"section": "0RGHOV(VWLPDWLRQ",
"sec_num": null
},
{
"text": "6WHS All possible word segmentations are generated using a Chinese lexicon containing 120,050 entries. The lexicon is only used for segmentation and there is no NE tag in it even if one word is PER, LOC or ORG. For example, \u00ebP (Beijing) is not tagged as LOC. 6WHS NE candidates are generated from any one or more segmented character strings and the corresponding generative prob ability for each candidate is computed using entity models described in Equation (4)-(7). 6WHS Viterbi search is used to select hypothesis with the highest probability as the best output. Furthermore, in order to identify nested named entities, two -pass Viterbi search is adopted. The inner Viterbi search is corresponding to Equation (6) and the outer one corresponding to Equation (1).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "0RGHOV(VWLPDWLRQ",
"sec_num": null
},
{
"text": "After the two-pass searches, the word segmentation and the named entities (including nested ones) can be obtained.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "0RGHOV(VWLPDWLRQ",
"sec_num": null
},
{
"text": "There are some problems with the framework of NE identification using the class-based LM. First, redundant candidates NEs are generated in the decoding process, which results in very large search space. The second problem is that data sparseness will seriously influence the performance. Finally, the abbreviation of NEs cannot be handled effectively. In the following three subsections, we provide solutions to the three problems mentioned above.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": ",PSURYHPHQW",
"sec_num": null
},
{
"text": "In order to overcome the redundant candidate generation problem, the heuristic information is introduced into the class-based LM. The following resources were used: (1) Chinese family name list, containing 373 entries (e.g. \u00f4 (Zhang), _ (Wang)); (2) transliterated name character list, containing 618 characters (e.g.( shi), 5 (dun)); and (3) ORG keyword list, containing 1,355 entries (e.g. \u00fb: (university), (\u00cc(corporation)).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "+HXULVWLF,QIRUPDWLRQ",
"sec_num": null
},
{
"text": "The heuristic information is used to constrain the generation of NE candidates. For PER (PN), only PER candidates beginning with the family name is considered. For PER (FN), a candidate is generated only if all its composing character belongs to the transliterated name character list. For ORG, a candidate is excluded if it does not contain one ORG keyword.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "+HXULVWLF,QIRUPDWLRQ",
"sec_num": null
},
{
"text": "Here, we do not utilize the LOC keyword to generate LOC candidate because of the fact that many LOC do not end with keywords.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "+HXULVWLF,QIRUPDWLRQ",
"sec_num": null
},
{
"text": "The cache entity model can address the data sparseness problem by adjusting the parameters continually as NE identification proceeds. The basic idea is to accumulate Chinese character or word n-gram so far appeared in the document and use them to create a local dynamic entity model such as",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&DFKH0RGHO",
"sec_num": null
},
{
"text": ") | ( 1 - L L ELFDFKH Z Z 3 and ) ( L XQLFDFKH Z 3",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&DFKH0RGHO",
"sec_num": null
},
{
"text": ". We can interpolate the cache entity model with the static entity LM ) ..",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&DFKH0RGHO",
"sec_num": null
},
{
"text": ". | ( 1 2 1 - - L L L VWDWLF Z Z Z Z 3 : Z Z Z _ Z 3 L L L FDFKH - - (8) ) .... | ( ) 1 ( ) | ( ) ( 1 1 2 1 1 2 1 - - - - + + = L L VWDWLF L L ELFDFKH L XQLFDFKH Z Z Z 3 Z Z 3 Z 3 O O O O where ] 1 , 0 [ , 2 1 \u00ce O O",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&DFKH0RGHO",
"sec_num": null
},
{
"text": "are interpolation weight that is determined on the held-out data set.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&DFKH0RGHO",
"sec_num": null
},
{
"text": "We found that many errors result from the occurrence of abbreviation of person, location, and organization. Therefore, different strategies are adopted to deal with abbreviations for different kinds of NEs. For PER, if Chinese surname is followed by the title, then this surname is tagged as PER. For example, \u00ba\u00f55 (President Zuo) is tagged as <PER>\u00ba</PER> \u00f55. For LOC, if at least two location abbreviations occur consecutive, the individual location abbreviation is tagged as LOC. For example, \u00b9 / \u00cf (Sino-Japan relation) is tagged as <LOC> </LOC><LOC>\u00b9</LOC> /\u00cf. For ORG, if organization abbreviation is followed by LOC, which is again followed by organization keyword, the three units are tagged as one ORG. For example, -\u00eb P \u00d6\u00a8(Chinese Communist Party Committee of Beijing) i s tagged as <ORG>-<LOC>\u00ebP</LOC> \u00d6 </ORG>. At present, we collected 112 organization abbreviations and 18 location abbreviations.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "'HDOLQJZLWK$EEUHYLDWLRQ",
"sec_num": null
},
{
"text": "We conduct evaluations in terms of precision (P) and recall (R). We also used the F-measure, which is defined as a weighted combination of precision and recall as Equation 11:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "([SHULPHQWV (YDOXDWLRQ0HWULF",
"sec_num": null
},
{
"text": "5 3 5 3 ) ++ = E E (11)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "([SHULPHQWV (YDOXDWLRQ0HWULF",
"sec_num": null
},
{
"text": "where E is the relative weight of precision and recall.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "([SHULPHQWV (YDOXDWLRQ0HWULF",
"sec_num": null
},
{
"text": "There are two differences between MET evaluation and ours. First, we include nested NE in our evaluation whereas MET does not. Second, in our evaluation, only NEs with correct boundary and type label are considered the correct identifications. In MET, the evaluation is somewhat flexible. For example, a NE may be identified partially correctly if the label is correct but the boundary is wrongly detected.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "([SHULPHQWV (YDOXDWLRQ0HWULF",
"sec_num": null
},
{
"text": "The training text corpus contains data from People's Daily (Jan.-Jun.1998). It contains 357,544 sentences (about 9,200,000 Chinese characters). This corpus includes 104,487 Chinese PER, 51,708 transliterated PER, 218,904 LOC, and 87,391 ORG. These data was obtained after this corpus was parsed with NLPWin.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "'DWD6HWV",
"sec_num": null
},
{
"text": "We built the wide coverage test data according to the guidelines 4 that are just same as those of 1999 IEER. The test set (as shown in Table 2 ) contains half a million Chinese characters; it is a balanced test set covering 11 domains. The test set contains 11,844 sentences, 49.84% of the sentences contain at least one NE. The number of characters in NE accounts for 8.448% in all Chinese characters.",
"cite_spans": [],
"ref_spans": [
{
"start": 135,
"end": 142,
"text": "Table 2",
"ref_id": null
}
],
"eq_spans": [],
"section": "'DWD6HWV",
"sec_num": null
},
{
"text": "We can see that the test data is much larger than the MET test data and IEER data ",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "'DWD6HWV",
"sec_num": null
},
{
"text": "The training data produced by NLPWin has some noise due to two reasons. First, the NE guideline used by NLPWin is different from the one we used. For example, in NLPWin, \u00eb P\u00d6(Beijing City) is tagged as <LOC>\u00ebP </LOC> \u00d6, whereas \u00ebP\u00d6 should be LOC in our definition. Second, there are some errors in NLPWin results. We utilized 18 rules to correct the frequent errors. The following shows some examples.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "7UDLQLQJ'DWD3UHSDUDWLRQ",
"sec_num": null
},
{
"text": "The Table 4 shows the quality of our training corpus. ",
"cite_spans": [],
"ref_spans": [
{
"start": 4,
"end": 11,
"text": "Table 4",
"ref_id": "TABREF4"
}
],
"eq_spans": [],
"section": "7UDLQLQJ'DWD3UHSDUDWLRQ",
"sec_num": null
},
{
"text": "We conduct incrementally the following f our experiments:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "([SHULPHQWV",
"sec_num": null
},
{
"text": "(1) Class-based LM, we view the results as baseline performance; (2) Integrating heuristic information into (1);",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "([SHULPHQWV",
"sec_num": null
},
{
"text": "(3) Integrating Cache-based LM with (2); (4) Integrating NE abbreviation processing with (3).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "([SHULPHQWV",
"sec_num": null
},
{
"text": "/1/RFDWLRQ.H\\ : /1 /1O /1Z",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "([SHULPHQWV",
"sec_num": null
},
{
"text": ":",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "([SHULPHQWV",
"sec_num": null
},
{
"text": "21 _>_\u00c5_\u00a9\u00ab \u00d1 : /1 &ODVVEDVHG/0%DVHOLQH",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "([SHULPHQWV",
"sec_num": null
},
{
"text": "Based on the basic class-based models estimated with the training data, we can get the baseline performance, as is shown in Table 5 .",
"cite_spans": [],
"ref_spans": [
{
"start": 124,
"end": 131,
"text": "Table 5",
"ref_id": "TABREF5"
}
],
"eq_spans": [],
"section": "([SHULPHQWV",
"sec_num": null
},
{
"text": "Comparing Table 4 and Table 5 , we found that the performance of baseline is better than the quality of training data. ,QWHJUDWLQJ+HXULVWLF,QIRUPDWLRQ",
"cite_spans": [],
"ref_spans": [
{
"start": 10,
"end": 17,
"text": "Table 4",
"ref_id": "TABREF4"
},
{
"start": 22,
"end": 29,
"text": "Table 5",
"ref_id": "TABREF5"
}
],
"eq_spans": [],
"section": "([SHULPHQWV",
"sec_num": null
},
{
"text": "In this part, we want to see the effects of using heuristic information. The results are shown in Table 6 . In experiments, we found that by integrating the heuristic information, we not only achieved more efficient decoding, but also obtained higher NE identification precision. For example, the precision of PER increases from 65.70% to 77.63%, and precision of ORG increases from 56.55% to 81.23%. The reason is that adopting heuristic information reduces the noise influence. However, we noticed that the recall of PER and LOC decreased a bit. There are two reasons. First, organization names without organization ending keywords were not marked as ORG. Second, Chinese names without surnames were also missed. ,QWHJUDWLQJ&DFKHEDVHG/0 Table 7 shows the evaluation results after cache-based LM was integrated. From Table 6 and Table 7 , we found that almost all the precision and recall of PER, LOC, ORG have obtained slight improvements. ,QWHJUDWLQJ ZLWK 1( $EEUHYLDWLRQ",
"cite_spans": [],
"ref_spans": [
{
"start": 98,
"end": 105,
"text": "Table 6",
"ref_id": "TABREF6"
},
{
"start": 739,
"end": 746,
"text": "Table 7",
"ref_id": "TABREF7"
},
{
"start": 818,
"end": 825,
"text": "Table 6",
"ref_id": "TABREF6"
},
{
"start": 830,
"end": 837,
"text": "Table 7",
"ref_id": "TABREF7"
}
],
"eq_spans": [],
"section": "([SHULPHQWV",
"sec_num": null
},
{
"text": "In this experiment, we integrated with NE abbreviation processing. As shown in Table 8 , the experiment result indicates that the recall of PER, LOC, ORG increased from 82.06%, 81.27%, 36.65% to 87.29%, 82.46%, 56.54%, respectively. ",
"cite_spans": [],
"ref_spans": [
{
"start": 79,
"end": 86,
"text": "Table 8",
"ref_id": "TABREF8"
}
],
"eq_spans": [],
"section": "3URFHVVLQJ",
"sec_num": null
},
{
"text": "From above data, we observed that (1) the class based SLM performs better than the training data automatically produced with the parser; (2) the distinct improvements is achieved by using heuristic information;",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "6XPPDU\\",
"sec_num": null
},
{
"text": "(3) Furthermore, our method of dealing with abbreviation increases the recall of NEs. In addition, the cache-based LM increases the performance not so much. The reason is as follows: The cache-based LM is based on the hypothesis that a word used in the recent past is much likely either to be used soon than its overall frequency in the language or a 3 -gram model would suggest (Kuhn, 1990) . However, we found that the same NE often vari es its morpheme in the same document. For example, the same NE -\u00eb P \u00d6\u00a8(Chinese Communist Party Committee of Beijing),\u00ebP \u00d6\u00a8(Committee of Beijing City), \u00d6( Committee) occur in order.",
"cite_spans": [
{
"start": 379,
"end": 391,
"text": "(Kuhn, 1990)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "6XPPDU\\",
"sec_num": null
},
{
"text": "Furthermore, we notice that the segmentation dictionary has an important impact on the performance of NE identification. We do not think it is better if more words are added into dictionary. For example, because \u00d1(Chinese) is in our dictionary, there is much possibility that \u00d1 (China) in \u00d1 is missed identified.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "6XPPDU\\",
"sec_num": null
},
{
"text": "We also evaluated on the MET2 test data and IEER test data. The results are shown in Table 9 . The results on MET2 are lower than the highest report of MUC7 (PER: Precision 66%, Recall 92%; LOC: Precision 89%, Recall 91%; ORG: Precision 89%, Recall 88%, http://www.itl.nist.gov). We speculate the reasons for this in the following. The main reason is that our class-based LM was estimated with a general domain corpus, which is quite different from the domain of MUC data. Moreover, we didn't use a NE dictionary. Another reason is that our NE definitions are slightly different from MET2. ",
"cite_spans": [],
"ref_spans": [
{
"start": 85,
"end": 93,
"text": "Table 9",
"ref_id": "TABREF9"
}
],
"eq_spans": [],
"section": "(YDOXDWLRQZLWK0(7DQG ,((57HVW'DWD",
"sec_num": null
},
{
"text": "In this research, Chinese word segmentation and NE identification has been integrated into a framework using class-based language models (LM). We adopted a hierarchical structure in ORG model so that the nested entities in organization names can be identified. Another characteristic is that our NE identification do not utilize NE dictionary when decoding. The evaluation on a large test set shows consistent improvements. The integration of heuristic information improves the precision and recall of our system. The cache-based LM increases the recall of NE identification to some extent. Moreover, some rules dealing with abbreviations of NEs have increased dramatically the performance. The precision of PER, LOC, ORG on the test set is 79.86%, 80.88%, 76.63%, respectively; and the recall is 87.29%, 82.46%, 56.54%, respectively.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&RQFOXVLRQV )XWXUHZRUN",
"sec_num": null
},
{
"text": "In our future work, we will be focusing more on NE coreference using language model. Second, we intend to extend our model to include the part-of-speech tagging model to improve the performance. At present, the class-based LM is based on the general domain and we may need to fine-tune the model for a specific domain. ACKNOWLEDGEMENT I would like to thank Ming Zhou, Jianfeng Gao, Changning Huang, Andi Wu, Hang Li and other colleagues from Microsoft Research for their help. And I want to thank especially Lei Zhang from Tsinghua University for his help in developing the ideas.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "&RQFOXVLRQV )XWXUHZRUN",
"sec_num": null
},
{
"text": "NLPWin system is a natural language processing system developed by Microsoft Research.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "The difference between IEER's guidelines and ours is that the nested person and location name in organization are tagged in our guidelines.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
}
],
"back_matter": [],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "A Maximum Entropy Approach to Named Entity Recognition",
"authors": [
{
"first": "",
"middle": [
"A"
],
"last": "Borthwick",
"suffix": ""
}
],
"year": 1999,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Borthwick. A. (1999) A Maximum Entropy Approach to Named Entity Recognition. PhD Dissertation",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "An algorithm that learns what's in a name",
"authors": [
{
"first": "D",
"middle": [],
"last": "Bikel",
"suffix": ""
},
{
"first": "R",
"middle": [],
"last": "Schwarta",
"suffix": ""
},
{
"first": "",
"middle": [
"R"
],
"last": "Weischedel",
"suffix": ""
}
],
"year": 1997,
"venue": "Machine Learning",
"volume": "34",
"issue": "",
"pages": "211--231",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Bikel D., Schwarta R., Weischedel. R. (1997) An algorithm that learns what's in a name. Machine Learning 34, pp. 211-231",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "Class-based n-gram models of natural language",
"authors": [
{
"first": "P",
"middle": [
"F"
],
"last": "Brown",
"suffix": ""
},
{
"first": "V",
"middle": [
"J"
],
"last": "Dellapietra",
"suffix": ""
},
{
"first": "P",
"middle": [
"V"
],
"last": "Desouza",
"suffix": ""
},
{
"first": "J",
"middle": [
"C"
],
"last": "Lai",
"suffix": ""
},
{
"first": "R",
"middle": [
"L"
],
"last": "Mercer",
"suffix": ""
}
],
"year": 1992,
"venue": "Computational Linguistics",
"volume": "18",
"issue": "4",
"pages": "468--479",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Brown, P. F., DellaPietra, V. J., deSouza, P. V., Lai, J. C., and Mercer, R. L. (1992). Class-based n-gram models of natural language. Computational Linguistics, 18(4):468--479.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "MUC-7 Named Entity Task Definition Version 3.5. Available by from ftp",
"authors": [
{
"first": "",
"middle": [
"N"
],
"last": "Chinchor",
"suffix": ""
}
],
"year": 1997,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Chinchor. N. (1997) MUC-7 Named Entity Task Definition Version 3.5. Available by from ftp.muc.saic.com/pub/MUC/MUC7-guidelines",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "Description of the NTU System Used for MET2",
"authors": [
{
"first": "H",
"middle": [
"H"
],
"last": "Chen",
"suffix": ""
},
{
"first": "Y",
"middle": [
"W"
],
"last": "Ding",
"suffix": ""
},
{
"first": "S",
"middle": [
"C"
],
"last": "Tsai",
"suffix": ""
},
{
"first": "G",
"middle": [
"W"
],
"last": "Bian",
"suffix": ""
}
],
"year": 1997,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Chen H.H., Ding Y.W., Tsai S.C. and Bian G.W. (1997) Description of the NTU System Used for MET2",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "Toward a unified Approach to Statistical Language Modeling for Chinese",
"authors": [
{
"first": "J",
"middle": [
"F"
],
"last": "Gao",
"suffix": ""
},
{
"first": "J",
"middle": [],
"last": "Goodman",
"suffix": ""
},
{
"first": "M",
"middle": [
"J"
],
"last": "Li",
"suffix": ""
},
{
"first": "K",
"middle": [
"F"
],
"last": "Lee",
"suffix": ""
}
],
"year": 2001,
"venue": "IEEE Transaction on Pattern Analysis and Machine Intelligence",
"volume": "12",
"issue": "6",
"pages": "570--583",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Gao J.F., Goodman J., Li M.J., Lee K.F. (2001) Toward a unified Approach to Statistical Language Modeling for Chinese. To appear in ACM Transaction on Asian Language Processing Kuhn R., Mori. R.D. (1990) A Cache-Based Natural Language Model for Speech Recognition. IEEE Transaction on Pattern Analysis and Machine Intelligence.Vol.12. No. 6. pp 570-583",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "Description of the LTG System Used",
"authors": [
{
"first": "A",
"middle": [],
"last": "Mikheev",
"suffix": ""
},
{
"first": "C",
"middle": [],
"last": "Grover",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Moens",
"suffix": ""
}
],
"year": 1997,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Mikheev A., Grover C. and Moens M. (1997) Description of the LTG System Used for MUC-7",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "A decision tree method for finding and classifying names in Japanese texts",
"authors": [
{
"first": "S",
"middle": [],
"last": "Sekine",
"suffix": ""
},
{
"first": "R",
"middle": [],
"last": "Grishman",
"suffix": ""
},
{
"first": "H",
"middle": [],
"last": "Shinou",
"suffix": ""
}
],
"year": 1998,
"venue": "Proceedings of the Sixth Workshop on Very Large Corpora",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Sekine S., Grishman R. and Shinou H. (1998), \"A decision tree method for finding and classifying names in Japanese texts\", Proceedings of the Sixth Workshop on Very Large Corpora, Canada",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "Description of the Kent Ridge Digital Labs System Used",
"authors": [
{
"first": "S",
"middle": [
"H"
],
"last": "Yu",
"suffix": ""
},
{
"first": "S",
"middle": [
"H"
],
"last": "Bai",
"suffix": ""
},
{
"first": "P",
"middle": [],
"last": "Wu",
"suffix": ""
}
],
"year": 1997,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Yu S.H., Bai S.H. and Wu P. (1997) Description of the Kent Ridge Digital Labs System Used for MUC-7",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "Study on Chinese Proofreading Oriented Language Modeling",
"authors": [
{
"first": "L",
"middle": [],
"last": "Zhang",
"suffix": ""
}
],
"year": 2001,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Zhang L. (2001) Study on Chinese Proofreading Oriented Language Modeling, PhD Dissertation",
"links": null
}
},
"ref_entries": {
"FIGREF0": {
"uris": null,
"num": null,
"text": "Example of Training Process'HFRGHU",
"type_str": "figure"
},
"TABREF3": {
"html": null,
"type_str": "table",
"content": "<table><tr><td>ID</td><td>Domain</td><td colspan=\"3\">Number of NE Tokens</td><td>Size</td></tr><tr><td/><td/><td colspan=\"2\">PER LOC</td><td>ORG</td><td>(byte)</td></tr><tr><td>1</td><td>Army</td><td>65</td><td>202</td><td>25</td><td>19k</td></tr><tr><td>2</td><td>Computer</td><td>75</td><td>156</td><td>171</td><td>59k</td></tr><tr><td>3</td><td>Culture</td><td>548</td><td>639</td><td>85</td><td>138k</td></tr><tr><td>4</td><td>Economy</td><td>160</td><td>824</td><td>363</td><td>108k</td></tr><tr><td>5</td><td>Entertainment</td><td>672</td><td>575</td><td>139</td><td>104k</td></tr><tr><td>6</td><td>Literature</td><td>464</td><td>707</td><td>122</td><td>96k</td></tr><tr><td>7</td><td>Nation</td><td>448</td><td>1193</td><td>250</td><td>101k</td></tr><tr><td>8</td><td>People</td><td>1147</td><td>912</td><td>403</td><td>116k</td></tr><tr><td>9</td><td>Politics</td><td>525</td><td>1148</td><td>218</td><td>122k</td></tr><tr><td>10</td><td>Science</td><td>155</td><td>204</td><td>87</td><td>60k</td></tr><tr><td>11</td><td>Sports</td><td>743</td><td>1198</td><td>628</td><td>114k</td></tr><tr><td/><td>Total</td><td colspan=\"2\">5002 7758</td><td colspan=\"2\">2491 1037k</td></tr></table>",
"num": null,
"text": "7DEOH: Statistics of Open-Test"
},
"TABREF4": {
"html": null,
"type_str": "table",
"content": "<table><tr><td/><td colspan=\"2\">Quality of Training Corpus</td><td/></tr><tr><td>NE</td><td>P (%)</td><td>R (%)</td><td>F (%)</td></tr><tr><td>PER</td><td>61.05</td><td>75.26</td><td>67.42</td></tr><tr><td>LOC</td><td>78.14</td><td>71.57</td><td>74.71</td></tr><tr><td>ORG</td><td>68.29</td><td>31.50</td><td>43.11</td></tr><tr><td>Total</td><td>70.07</td><td>66.08</td><td>68.02</td></tr></table>",
"num": null,
"text": ""
},
"TABREF5": {
"html": null,
"type_str": "table",
"content": "<table><tr><td/><td colspan=\"2\">Baseline Performance</td><td/></tr><tr><td>NE</td><td>P (%)</td><td>R (%)</td><td>F (%)</td></tr><tr><td>PER</td><td>65.70</td><td>84.37</td><td>73.87</td></tr><tr><td>LOC</td><td>82.73</td><td>76.03</td><td>79.24</td></tr><tr><td>ORG</td><td>56.55</td><td>38.56</td><td>45.86</td></tr><tr><td>Total</td><td>72.61</td><td>72.44</td><td>72.53</td></tr></table>",
"num": null,
"text": ""
},
"TABREF6": {
"html": null,
"type_str": "table",
"content": "<table><tr><td/><td colspan=\"3\">Results of Heuristic Information Integrated</td></tr><tr><td colspan=\"2\">into the Class-based LM</td><td/><td/></tr><tr><td>NE</td><td>P (%)</td><td>R (%)</td><td>F (%)</td></tr><tr><td>PER</td><td>77.63</td><td>80.89</td><td>79.23</td></tr><tr><td>LOC</td><td>80.05</td><td>80.80</td><td>80.42</td></tr><tr><td>ORG</td><td>81.23</td><td>36.65</td><td>50.51</td></tr><tr><td>Total</td><td>79.26</td><td>73.41</td><td>76.23</td></tr></table>",
"num": null,
"text": ""
},
"TABREF7": {
"html": null,
"type_str": "table",
"content": "<table><tr><td/><td colspan=\"2\">Results of our system</td><td/></tr><tr><td>NE</td><td>P (%)</td><td>R (%)</td><td>F (%)</td></tr><tr><td>PER</td><td>79.12</td><td>82.06</td><td>80.57</td></tr><tr><td>LOC</td><td>80.11</td><td>81.27</td><td>80.69</td></tr><tr><td>ORG</td><td>79.71</td><td>39.89</td><td>53.17</td></tr><tr><td>Total</td><td>79.72</td><td>74.58</td><td>77.06</td></tr></table>",
"num": null,
"text": ""
},
"TABREF8": {
"html": null,
"type_str": "table",
"content": "<table><tr><td/><td colspan=\"2\">Results of our system</td><td/></tr><tr><td>NE</td><td>P (%)</td><td>R (%)</td><td>F (%)</td></tr><tr><td>PER</td><td>79.86</td><td>87.29</td><td>83.41</td></tr><tr><td>LOC</td><td>80.88</td><td>82.46</td><td>81.66</td></tr><tr><td>ORG</td><td>76.63</td><td>56.54</td><td>65.07</td></tr><tr><td>Total</td><td>79.99</td><td>79.68</td><td>79.83</td></tr></table>",
"num": null,
"text": ""
},
"TABREF9": {
"html": null,
"type_str": "table",
"content": "<table><tr><td>NE</td><td colspan=\"2\">MET2 Data</td><td/><td colspan=\"2\">IEER Data</td></tr><tr><td/><td>P</td><td>R</td><td>F</td><td>P</td><td>R</td><td>F</td></tr><tr><td/><td>(%)</td><td>(%)</td><td>(%)</td><td>(%)</td><td>(%)</td><td>(%)</td></tr><tr><td>PER</td><td>65.86</td><td/><td/><td/><td/></tr></table>",
"num": null,
"text": "Results on MET2 and IEER 94.25 77.54 79.38 84.43 81.83 LOC 77.42 89.60 83.07 79.09 80.18 79.63 ORG 88.47 75.33 81.38 88.03 62.30 72.96 Total 77.89 86.09 81.79 80.82 76.78 78.75"
}
}
}
} |