File size: 67,183 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 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 |
{
"paper_id": "W10-0202",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T05:04:30.493244Z"
},
"title": "Emotion Detection in Email Customer Care",
"authors": [
{
"first": "Narendra",
"middle": [],
"last": "Gupta",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "AT&T Labs -Research, Inc. Florham Park",
"location": {
"postCode": "07932",
"region": "NJ",
"country": "USA"
}
},
"email": "[email protected]"
},
{
"first": "Mazin",
"middle": [],
"last": "Gilbert",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "AT&T Labs -Research, Inc. Florham Park",
"location": {
"postCode": "07932",
"region": "NJ",
"country": "USA"
}
},
"email": ""
},
{
"first": "Giuseppe",
"middle": [
"Di"
],
"last": "Fabbrizio",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "AT&T Labs -Research, Inc. Florham Park",
"location": {
"postCode": "07932",
"region": "NJ",
"country": "USA"
}
},
"email": ""
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "Prompt and knowledgeable responses to customers' emails are critical in maximizing customer satisfaction. Such emails often contain complaints about unfair treatment due to negligence, incompetence, rigid protocols, unfriendly systems, and unresponsive personnel. In this paper, we refer to these emails as emotional emails. They provide valuable feedback to improve contact center processes and customer care, as well as, to enhance customer retention. This paper describes a method for extracting salient features and identifying emotional emails in customer care. Salient features reflect customer frustration, dissatisfaction with the business, and threats to either leave, take legal action and/or report to authorities. Compared to a baseline system using word ngrams, our proposed approach with salient features resulted in a 20% absolute Fmeasure improvement.",
"pdf_parse": {
"paper_id": "W10-0202",
"_pdf_hash": "",
"abstract": [
{
"text": "Prompt and knowledgeable responses to customers' emails are critical in maximizing customer satisfaction. Such emails often contain complaints about unfair treatment due to negligence, incompetence, rigid protocols, unfriendly systems, and unresponsive personnel. In this paper, we refer to these emails as emotional emails. They provide valuable feedback to improve contact center processes and customer care, as well as, to enhance customer retention. This paper describes a method for extracting salient features and identifying emotional emails in customer care. Salient features reflect customer frustration, dissatisfaction with the business, and threats to either leave, take legal action and/or report to authorities. Compared to a baseline system using word ngrams, our proposed approach with salient features resulted in a 20% absolute Fmeasure improvement.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "Emails are becoming the preferred communication channel for customer service. For customers, it is a way to avoid long hold times on call centers phone calls and to keep a record of the information exchanges with the business. For businesses, it offers an opportunity to best utilize customer service representatives by evenly distributing the work load over time, and for representatives, it allows time to research the issue and respond to the customers in a manner consistent with business policies. Businesses can further exploit the offline nature of this channel by automatically routing the emails involving critical issues to specialized representatives. Besides concerns related to products and services, businesses ensure that emails complaining about unfair treatment due to negligence, incompetence, rigid protocols and unfriendly systems, are always handled with care. Such emails, referred to as emotional emails, are critical to reduce the churn i.e., retaining customers who otherwise would have taken their business elsewhere, and, at the same time, they are a valuable source of information for improving business processes.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "In recurring service oriented businesses, a large number of customer emails may contain routine complaints. While such complaints are important and are addressed by customer service representatives, our purpose here is to identify emotional emails where severity of the complaints and customer dissatisfaction are relatively high. Emotional emails may contain abusive and probably emotionally charged language, but we are mainly interested in identifying messages where, in addition to the flames, the customer includes a concrete description of the problem experienced with the company providing the service. In the context of customer service, customers express their concerns in many ways. Sometimes they convey a negative emotional component articulated by phrases like disgusted and you suck. In other cases, there is a minimum emotional involvement by enumerating factual sentences such as you overcharged, or take my business elsewhere. In many cases, both the emotional and factual components are actually present. In this work, we have identified eight dif-ferent ways that customers use to express their emotions in emails. Throughout this paper, these ways will be referred to as Salient Features. We cast the identification of emotional email as a text classification problem, and show that using salient features we can significantly improve the identification accuracy. Compared to a baseline system which uses Boosting (Schapire, 1999) withnword n-grams features, our proposed system using salient features resulted in improvement in f-measure from 0.52 to 0.72.",
"cite_spans": [
{
"start": 1434,
"end": 1450,
"text": "(Schapire, 1999)",
"ref_id": "BIBREF17"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "In section 2, we provide a summary of previous work and its relationship with our contribution. In section 3, we describe our method for emotion detection and extraction of salient features. A series of experiments demonstrating improvement in classification performance is presented in section 4. We conclude the paper by highlighting the main contribution of this work in section 5.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Extensive work has been done on emotion detection. In the context of human-computer dialogs, although richer features including acoustic and intonation are available, there is a general consensus (Litman and Forbes-Riley, 2004b; Lee and Narayanan, 2005) about the use of lexical features to significantly improve the accuracy of emotion detection.",
"cite_spans": [
{
"start": 229,
"end": 253,
"text": "Lee and Narayanan, 2005)",
"ref_id": "BIBREF11"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Previous Work",
"sec_num": "2"
},
{
"text": "Research has also been done in predicting basic emotions (also referred to as affects) within text Liu et al., 2003) . To render speech with prosodic contour conveying the emotional content of the text, one of 6 types of human emotions (e.g., angry, disgusted, fearful, happy, sad, and surprised) are identified for each sentence in the running text. Deducing such emotions from lexical constructs is a hard problem evidenced by little agreement among humans. A Kappa value of 0.24-0.51 was shown in . Liu et al. (2003) have argued that the absence of affect laden surface features i.e., key words, from the text does not imply absence of emotions, therefore they have relied more on common-sense knowledge. Instead of deducing types emotions in each sentence, we are interested in knowing if the entire email is emotional or not. Additionally we are also interested in the intensity and the cause of those emotions.",
"cite_spans": [
{
"start": 99,
"end": 116,
"text": "Liu et al., 2003)",
"ref_id": "BIBREF15"
},
{
"start": 502,
"end": 519,
"text": "Liu et al. (2003)",
"ref_id": "BIBREF15"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Previous Work",
"sec_num": "2"
},
{
"text": "There is also a body of work in areas of creating Semantic Orientation (SO) dictionaries (Hatzivassiloglou and McKeown, 1997; Turney and Littman, 2003; Esuli and Sebastiani, 2005) and their use in identifying emotions laden sentences and polarity (Yu and Hatzivassiloglou, 2003; Kim and Hovy, 2004; Hu and Liu, 2004) of those emotions. While such dictionaries provide a useful starting point, their use alone does not yield satisfactory results. In Wilson et al. (2005) , classification of phrases containing positive, negative or neutral emotions is discussed. For this problem they show high agreement among human annotators (Kappa of 0.84). They also show that labeling phrases as positive, negative or neutral only on the basis of presence of key word from such dictionaries yields a classification accuracy of 48%. An obvious reason for this poor performance is that semantic orientations of words are context dependent.",
"cite_spans": [
{
"start": 89,
"end": 125,
"text": "(Hatzivassiloglou and McKeown, 1997;",
"ref_id": "BIBREF7"
},
{
"start": 126,
"end": 151,
"text": "Turney and Littman, 2003;",
"ref_id": "BIBREF20"
},
{
"start": 152,
"end": 179,
"text": "Esuli and Sebastiani, 2005)",
"ref_id": "BIBREF4"
},
{
"start": 247,
"end": 278,
"text": "(Yu and Hatzivassiloglou, 2003;",
"ref_id": "BIBREF22"
},
{
"start": 279,
"end": 298,
"text": "Kim and Hovy, 2004;",
"ref_id": "BIBREF10"
},
{
"start": 299,
"end": 316,
"text": "Hu and Liu, 2004)",
"ref_id": "BIBREF9"
},
{
"start": 449,
"end": 469,
"text": "Wilson et al. (2005)",
"ref_id": "BIBREF21"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Previous Work",
"sec_num": "2"
},
{
"text": "Works reported in Wilson et al. (2005) ; Pang et al. (2002) and Dave et al. (2003) have attempted to mitigate this problem by using supervised methods. They report classification results using a number of different sets of features, including unigram word features. Wilson et al. (2005) reports an improvement (63% to 65.7% accuracy) in performance by using a host of features extracted from syntactic dependencies. Similarly, Gamon (2004) shows that the use of deep semantic features along with word unigrams improve performances. Pang et al. (2002) and Dave et al. (2003) on the other hand confirmed that word unigrams provide the best classification results. This is in line with our experience as well and could be due to sparseness of the data. We also used supervised methods to predict emotional emails. To train predictive models we used word ngrams (uni-, bi-and tri-grams) and a number of binary features indicating the presence of words/phrases from specific dictionaries. Spertus (1997) discusses a system called Smoky which recognizes hostile messages and is quite similar to our work. While Smoky is interested in identifying messages that contain flames, our research on emotional emails looks deeper to discover the reasons for such flames. Besides word unigrams, Smoky uses rules to derive additional features for classification. These features are intended to capture different manifestations of the flames. Simi-larly, in our work we also use rules (in our case implemented as table look-up) to derive additional features of emotional emails.",
"cite_spans": [
{
"start": 18,
"end": 38,
"text": "Wilson et al. (2005)",
"ref_id": "BIBREF21"
},
{
"start": 41,
"end": 59,
"text": "Pang et al. (2002)",
"ref_id": "BIBREF16"
},
{
"start": 64,
"end": 82,
"text": "Dave et al. (2003)",
"ref_id": "BIBREF3"
},
{
"start": 266,
"end": 286,
"text": "Wilson et al. (2005)",
"ref_id": "BIBREF21"
},
{
"start": 427,
"end": 439,
"text": "Gamon (2004)",
"ref_id": "BIBREF5"
},
{
"start": 532,
"end": 550,
"text": "Pang et al. (2002)",
"ref_id": "BIBREF16"
},
{
"start": 555,
"end": 573,
"text": "Dave et al. (2003)",
"ref_id": "BIBREF3"
},
{
"start": 984,
"end": 998,
"text": "Spertus (1997)",
"ref_id": "BIBREF19"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Previous Work",
"sec_num": "2"
},
{
"text": "We use supervised machine learning techniques to detect emotional emails. In particular, our emotion detector is a statistical classifier model trained using hand labeled training examples. For each example, a set of salient features is extracted. The major components of our system are described below.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Emotion detection in emails",
"sec_num": "3"
},
{
"text": "For detecting emotional emails we used Boostexter as text classification. Our choice of machine learning algorithm was not strategic and we have no reason to believe that SVMs or maximum entropybased classifiers will not perform equally well. Boostexter, which is based on the boosting family of algorithms, was first proposed by Schapire (1999) . It has been applied successfully to numerous text classification applications (Gupta et al., 2005) at AT&T. Boosting builds a highly accurate classifier by combining many \"weak\" base classifiers, each one of which may only be moderately accurate. Boosting constructs the collection of base classifiers iteratively. On each iteration t, the boosting algorithm supplies the base learner weighted training data and the base learner generates a base classifier h t . Set of nonnegative weights w t encode how important it is that h t correctly classifies each email. Generally, emails that were most often misclassified by the preceding base classifiers will be given the most weight so as to force the base learner to focus on the \"hardest\" examples. As described in Schapire and Singer (1999) , Boostexter uses confidence rated base classifiers h that for every example x (in our case it is the customer emails) output a real number h(x) whose sign (-1 or +1) is interpreted as a prediction(+1 indicates emotional email), and whose magnitude |h(x)| is a measure of \"confidence.\" The output of the final classifier f is f (x) = T t=1 h t (x), i.e., the sum of confidence of all classifiers h t . The real-valued predictions of the final classifier f can be mapped onto a confidence value between 0 and 1 by a logistic function;",
"cite_spans": [
{
"start": 330,
"end": 345,
"text": "Schapire (1999)",
"ref_id": "BIBREF17"
},
{
"start": 426,
"end": 446,
"text": "(Gupta et al., 2005)",
"ref_id": "BIBREF6"
},
{
"start": 1112,
"end": 1138,
"text": "Schapire and Singer (1999)",
"ref_id": "BIBREF18"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Classifier",
"sec_num": "3.1"
},
{
"text": "conf (x = emotional email) = 1 1 + e \u2212f (x) .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Classifier",
"sec_num": "3.1"
},
{
"text": "The learning procedure in boosting minimizes the negative conditional log likelihood of the training data under this model, namely:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Classifier",
"sec_num": "3.1"
},
{
"text": "i ln(1 + e \u2212y i f (x i ) ).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Classifier",
"sec_num": "3.1"
},
{
"text": "Here i iterates over all training examples and y i is the label of ith example.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Classifier",
"sec_num": "3.1"
},
{
"text": "Emotional emails are a reaction to perceived excessive loss of time and/or money by customers. Expressions of such reactions in emails are salient features of emotional emails. For our data we have identified the eight features listed below. While many of these features are of general nature and can be present in most customer service related emotional emails, in this paper we make no claims about their completeness.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Feature extraction",
"sec_num": "3.2"
},
{
"text": "itly expressing customers affective states by phrases like it upsets me, I am frustrated;",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Expression of negative emotions: Explic-",
"sec_num": "1."
},
{
"text": "2. Expression of negative opinions about the company: by evaluative expressions like dishonest dealings, disrespectful.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Expression of negative emotions: Explic-",
"sec_num": "1."
},
{
"text": "These could also be insulting expressions like stink, suck, idiots;",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Expression of negative emotions: Explic-",
"sec_num": "1."
},
{
"text": "3. Threats to take their business elsewhere: by expression like business elsewhere, look for another provider. These expressions are neither emotional or evaluative;",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Expression of negative emotions: Explic-",
"sec_num": "1."
},
{
"text": "4. Threats to report to authorities: federal agencies, consumer protection. These are domain dependent names of agencies. The mere presence of such names implies customer threat;",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Expression of negative emotions: Explic-",
"sec_num": "1."
},
{
"text": "5. Threats to take legal action:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Expression of negative emotions: Explic-",
"sec_num": "1."
},
{
"text": "seek retribution, lawsuit.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Expression of negative emotions: Explic-",
"sec_num": "1."
},
{
"text": "These expressions may also not be emotional or evaluative in nature; 6. Justification about why they should have been treated better. A common way to do this is to say things like long time customer, loyal customer, etc. Semantic orientations of most phrases used to express this feature are positive;",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Expression of negative emotions: Explic-",
"sec_num": "1."
},
{
"text": "7. Disassociate themselves from the company, by using phrases like you people, your service representative, etc. These are analogous to rule class \"Noun Phrases used as Appositions\" in Spertus (1997) .",
"cite_spans": [
{
"start": 185,
"end": 199,
"text": "Spertus (1997)",
"ref_id": "BIBREF19"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Expression of negative emotions: Explic-",
"sec_num": "1."
},
{
"text": "8. State what was done wrong to them: grossly overcharged, on hold for hours, etc. These phrases may have negative or neutral semantic orientations.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Expression of negative emotions: Explic-",
"sec_num": "1."
},
{
"text": "In addition to the word unigrams, salient features of emotional emails are also used for training/testing the emotional email classifier. While labeling the training data, labelers look for salient features within the email and also the severity of the loss perceived by the customer. For example, email 1 in Fig. 1 is labeled as emotional because customer perception of loss is severe to the point that the customer may cancel the service. On the other hand, email 2 is not emotional because customer perceived loss is not severe to the point of service cancellation. This customer would be satisfied in this instant if he/she receives the requested information in a timely fashion.",
"cite_spans": [],
"ref_spans": [
{
"start": 309,
"end": 315,
"text": "Fig. 1",
"ref_id": "FIGREF0"
}
],
"eq_spans": [],
"section": "Expression of negative emotions: Explic-",
"sec_num": "1."
},
{
"text": "To extract salient features from an email, eight separate lists of phrases customers use to express each of the salient features were manually created. These lists were extracted from the training data and can be considered as basic rules that identify emotional emails. In the labeling guide for critical emails labelers were instructed to look for salient features in the email and keep a list of encountered phrases. We further enriched these lists by: a) using general knowledge of English, we added variations to existing phrases and b) searching a large body of email text (different from testing) for different phrases in which key words from known phrases participated. For example from the known phrase lied to we used the word lied and found a phrase blatantly lied. Using these lists we extracted eight binary salient features for each email, indicating presence/absence of phrases from the corresponding list in the email. ",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Expression of negative emotions: Explic-",
"sec_num": "1."
},
{
"text": "We performed several experiments to compare the performance of our emotional email classifier with that using a ngram based text classifier. For these experiments we labeled 620 emails as training examples and 457 emails as test examples. Training examples were labeled independently by two different labelers 1 with relatively high degree of agreement among them. Kappa (Cohen, 1960) value of 0.814 was observed versus 0.5-0.7 reported for emotion labeling tasks Litman and Forbes-Riley, 2004a ). Because of the relatively high agreement among these labelers, with different back ground, we did not feel the need to check the agreement among more than 2 labelers. Due to the limited size of the training data we used cross validation (leave-one-out) technique on the test set to evaluate outcomes of different experiments. In this round robin approach, each example from the test set is tested using a model trained on all remaining 1076 (620 plus 456) examples. Test results on all 457 test examples are averaged.",
"cite_spans": [
{
"start": 371,
"end": 384,
"text": "(Cohen, 1960)",
"ref_id": "BIBREF2"
},
{
"start": 464,
"end": 494,
"text": "Litman and Forbes-Riley, 2004a",
"ref_id": "BIBREF12"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Experiments and evaluation",
"sec_num": "4"
},
{
"text": "Throughout all of our experiments, we computed the classification accuracy of detecting emotional emails using precision, recall and F-measure. Notice for our test data a classifier with majority vote has a classification accuracy of 87%, but since none of the emotional emails are identified, recall and Fmeasure are both zero. On the other hand, a classifier which generates many more false positives for each true positive, will have a lower classification accuracy but a higher (non-zero) F-measure than the majority vote classifier. Fig. 2 shows precision/recall curves for different experiments. The black circles represent the operating point corresponding to the best F-measure for each curve. Actual values of these points are provided in Table 2 .",
"cite_spans": [],
"ref_spans": [
{
"start": 538,
"end": 544,
"text": "Fig. 2",
"ref_id": "FIGREF1"
},
{
"start": 748,
"end": 755,
"text": "Table 2",
"ref_id": "TABREF4"
}
],
"eq_spans": [],
"section": "Experiments and evaluation",
"sec_num": "4"
},
{
"text": "As a baseline experiment we used word ngram features to train a classifier model. The graph labeled as \"ngram features\" in Fig. 2 shows the performance of this classifier. The best F-measure in this case is only 0.52. Obviously this low performance can be attributed to the small training set and the large feature space formed by word ngrams. ",
"cite_spans": [],
"ref_spans": [
{
"start": 123,
"end": 129,
"text": "Fig. 2",
"ref_id": "FIGREF1"
}
],
"eq_spans": [],
"section": "Experiments and evaluation",
"sec_num": "4"
},
{
"text": "The baseline system was compared with a similar system using salient features. First, we used a simple classification rule that we formulated by looking at the training data. According to this rule, if an email contained three or more salient features it was classified as an emotional email. We classified the test data using this rule and obtained and an Fmeasure of 0.68 (see row labeled as \u2265 3 in Table 2 ).",
"cite_spans": [],
"ref_spans": [
{
"start": 401,
"end": 408,
"text": "Table 2",
"ref_id": "TABREF4"
}
],
"eq_spans": [],
"section": "Salient features",
"sec_num": "4.1"
},
{
"text": "Since no confidence thresholding can be used with the deterministic rule, its performance is indicated by a single point marked by the gray circle in Fig. 2 . This result clearly demonstrates high utility of our salient features. To verify that the salient features threshold count of 3 used in our simple classification rule is the best, we also evaluated the performance of the rule for the salient features with threshold count of 2 and 4 (row labeled as \u2265 2 and \u2265 4 in Table 2 ). In our next set experiments, we trained a classifier model using salient features alone and with word ngrams. Corresponding cross validation results on the test data are annotated in Table 2 and in Fig. 2 as \"Salient Features\" and \"N-grams & Salient Features\", respectively. Incremental improvement in best F-measure clearly shows: a) BoosTexter is able to learn better rules than the simple rule of identifying three or more salient features. b) Even though salient features provide a significant improvement in performance, there is still discriminative information in ngram features. A direct consequence of the second observation is that the detection accuracy can be further improved by extending/refining the phrase lists and/or by using more labeled data so that to exploit the discriminative information in the word ngram features.",
"cite_spans": [],
"ref_spans": [
{
"start": 150,
"end": 156,
"text": "Fig. 2",
"ref_id": "FIGREF1"
},
{
"start": 473,
"end": 480,
"text": "Table 2",
"ref_id": "TABREF4"
},
{
"start": 667,
"end": 674,
"text": "Table 2",
"ref_id": "TABREF4"
},
{
"start": 682,
"end": 688,
"text": "Fig. 2",
"ref_id": "FIGREF1"
}
],
"eq_spans": [],
"section": "Salient features",
"sec_num": "4.1"
},
{
"text": "Salient Features of emotional emails are the consequence of our knowledge of how customers react to their excessive loss. To empirically demonstrate that eight different salient features used in identification of emotional emails do provide complementary evidence, we randomly distributed the phrases in eight lists. We then used them to extract eight binary features in the same manner as before. Best F-measure for this experiment is shown in the last row of Table 2 , and labeled as \"N-gram & Random Features\". Degradation in performance of this experiment clearly demonstrates that salient features used by us provide complimentary and not redundant information.",
"cite_spans": [],
"ref_spans": [
{
"start": 461,
"end": 468,
"text": "Table 2",
"ref_id": "TABREF4"
}
],
"eq_spans": [],
"section": "Salient features",
"sec_num": "4.1"
},
{
"text": "Customer emails complaining about unfair treatment are often emotional and are critical for businesses. They provide valuable feedback for improving business processes and coaching agents. Furthermore careful handling of such emails helps to improve customer retention. In this paper, we presented a method for emotional email identification. We introduced the notion of salient features for emotional emails, and demonstrated high agreement among two labelers in detecting emotional emails. We also demonstrated that extracting salient features from the email text and using them to train a classifier model can significantly improve identification accuracy. Compared to a baseline classifier which uses only the word ngrams features, the addition of the salient features improved the F-measure from 0.52 to 0.72. Our current research is focused on improving the salient feature extraction process.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusions",
"sec_num": "5"
},
{
"text": "More specifically by leveraging publically available Semantic orientation dictionaries, and by enriching our dictionaries using phrases extracted from a large corpus by matching syntactic patterns of some seed phrases.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusions",
"sec_num": "5"
},
{
"text": "One of the labeler was one of the authors of this paper and other had linguistic back ground.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
}
],
"back_matter": [],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "Emotional sequencing and development in fairy tales",
"authors": [
{
"first": "Cecilia",
"middle": [],
"last": "Alm",
"suffix": ""
},
{
"first": "Richard",
"middle": [],
"last": "Sproat",
"suffix": ""
}
],
"year": 2005,
"venue": "Proceedings of the First International Conference on Affective Computing and Intelligent Interaction",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Alm, Cecilia and Richard Sproat. 2005. Emotional sequencing and development in fairy tales. In Proceedings of the First International Conference on Affective Computing and Intelligent Interac- tion.",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "Emotions from text: machine learning for text-based emotion prediction",
"authors": [
{
"first": "Cecilia",
"middle": [],
"last": "Alm",
"suffix": ""
},
{
"first": "Dan",
"middle": [],
"last": "Ovesdotter",
"suffix": ""
},
{
"first": "Richard",
"middle": [],
"last": "Roth",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Sproat",
"suffix": ""
}
],
"year": 2005,
"venue": "HLT '05: Proceedings of the conference on Human Language Technology and Empirical Methods in Natural Language Processing",
"volume": "",
"issue": "",
"pages": "579--586",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Alm, Cecilia Ovesdotter, Dan Roth, and Richard Sproat. 2005. Emotions from text: machine learning for text-based emotion prediction. In HLT '05: Proceedings of the conference on Hu- man Language Technology and Empirical Meth- ods in Natural Language Processing. Association for Computational Linguistics, Morristown, NJ, USA, pages 579-586.",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "A coefficient of agreement for nominal scales",
"authors": [
{
"first": "J",
"middle": [],
"last": "Cohen",
"suffix": ""
}
],
"year": 1960,
"venue": "Educational and Psychological Measurement",
"volume": "20",
"issue": "1",
"pages": "37--46",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Cohen, J. 1960. A coefficient of agreement for nom- inal scales. Educational and Psychological Mea- surement 20(1):37-46.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "Mining the peanut gallery: Opinion extraction and semantic classification of product reviews",
"authors": [
{
"first": "Kushal",
"middle": [],
"last": "Dave",
"suffix": ""
},
{
"first": "Steve",
"middle": [],
"last": "Lawrence",
"suffix": ""
},
{
"first": "David",
"middle": [
"M"
],
"last": "Pennock",
"suffix": ""
}
],
"year": 2003,
"venue": "Proceedings of WWW",
"volume": "",
"issue": "",
"pages": "519--528",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Dave, Kushal, Steve Lawrence, and David M. Pen- nock. 2003. Mining the peanut gallery: Opinion extraction and semantic classification of product reviews. In Proceedings of WWW. pages 519- 528.",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "Determining the semantic orientation of terms through gloss classificaion",
"authors": [
{
"first": "A",
"middle": [],
"last": "Esuli",
"suffix": ""
},
{
"first": "F",
"middle": [],
"last": "Sebastiani",
"suffix": ""
}
],
"year": 2005,
"venue": "14th ACM International Conference on Information and Knowledge Management",
"volume": "",
"issue": "",
"pages": "617--624",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Esuli, A. and F. Sebastiani. 2005. Determin- ing the semantic orientation of terms through gloss classificaion. In Proceedings of CIKM-05, 14th ACM International Conference on Informa- tion and Knowledge Management. Bremen, DE., pages 617-624.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "Sentiment classification on customer feedback data: Noisy data large feature vectors and the role of linguistic analysis",
"authors": [
{
"first": "M",
"middle": [],
"last": "Gamon",
"suffix": ""
}
],
"year": 2004,
"venue": "Proceedings of COLING",
"volume": "",
"issue": "",
"pages": "841--847",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Gamon, M. 2004. Sentiment classification on cus- tomer feedback data: Noisy data large feature vectors and the role of linguistic analysis. In Pro- ceedings of COLING 2004. Geneva, Switzerland, pages 841-847.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "The AT&T Spoken Language Understanding System",
"authors": [
{
"first": "Narendra",
"middle": [],
"last": "Gupta",
"suffix": ""
},
{
"first": "Gokhan",
"middle": [],
"last": "Tur",
"suffix": ""
},
{
"first": "Dilek",
"middle": [],
"last": "Hakkani-T\u00fcr",
"suffix": ""
},
{
"first": "Srinivas",
"middle": [],
"last": "Banglore",
"suffix": ""
},
{
"first": "Giuseppe",
"middle": [],
"last": "Riccardi",
"suffix": ""
},
{
"first": "Mazin",
"middle": [],
"last": "Rahim",
"suffix": ""
}
],
"year": 2005,
"venue": "IEEE Transactions on Speech and Audio Processing",
"volume": "14",
"issue": "1",
"pages": "213--222",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Gupta, Narendra, Gokhan Tur, Dilek Hakkani-T\u00fcr, Srinivas Banglore, Giuseppe Riccardi, and Mazin Rahim. 2005. The AT&T Spoken Language Understanding System. IEEE Transactions on Speech and Audio Processing 14(1):213-222.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "Predicting the semantic orientation of ad-15",
"authors": [
{
"first": "Vasileios",
"middle": [],
"last": "Hatzivassiloglou",
"suffix": ""
},
{
"first": "Kathleen",
"middle": [],
"last": "Mckeown",
"suffix": ""
}
],
"year": 1997,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Hatzivassiloglou, Vasileios and Kathleen McKeown. 1997. Predicting the semantic orientation of ad- 15",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "Proceedings of the Joint ACL/EACL Conference",
"authors": [],
"year": null,
"venue": "",
"volume": "",
"issue": "",
"pages": "174--181",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "jectives. In Proceedings of the Joint ACL/EACL Conference. pages 174-181.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "Mining and summarizing customer reviews",
"authors": [
{
"first": "Minqing",
"middle": [],
"last": "Hu",
"suffix": ""
},
{
"first": "Bing",
"middle": [],
"last": "Liu",
"suffix": ""
}
],
"year": 2004,
"venue": "Proceedings of the ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD)",
"volume": "",
"issue": "",
"pages": "168--177",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Hu, Minqing and Bing Liu. 2004. Mining and sum- marizing customer reviews. In Proceedings of the ACM SIGKDD Conference on Knowledge Dis- covery and Data Mining (KDD). pages 168-177.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "Determining the sentiment of opinions",
"authors": [
{
"first": "Soo-Min",
"middle": [],
"last": "Kim",
"suffix": ""
},
{
"first": "Eduard",
"middle": [],
"last": "Hovy",
"suffix": ""
}
],
"year": 2004,
"venue": "Proceedings of the International Conference on Computational Linguistics (COLING)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Kim, Soo-Min and Eduard Hovy. 2004. Determin- ing the sentiment of opinions. In Proceedings of the International Conference on Computational Linguistics (COLING).",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "Toward detecting emotions in spoken dialogs",
"authors": [
{
"first": "Chul",
"middle": [],
"last": "Lee",
"suffix": ""
},
{
"first": "Shrikanth",
"middle": [
"S"
],
"last": "Min",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Narayanan",
"suffix": ""
}
],
"year": 2005,
"venue": "IEEE Transactions on Speech and Audio Processing",
"volume": "13",
"issue": "2",
"pages": "293--303",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Lee, Chul Min and Shrikanth S. Narayanan. 2005. Toward detecting emotions in spoken dialogs. IEEE Transactions on Speech and Audio Process- ing 13(2):293-303.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "Annotating student emotional states in spoken tutoring dialogues",
"authors": [
{
"first": "D",
"middle": [],
"last": "Litman",
"suffix": ""
},
{
"first": "K",
"middle": [],
"last": "Forbes-Riley",
"suffix": ""
}
],
"year": 2004,
"venue": "Proceedings of the 5th SIGdial Workshop on Discourse and Dialogue (SIGdial)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Litman, D. and K. Forbes-Riley. 2004a. Annotat- ing student emotional states in spoken tutoring dialogues. In Proceedings of the 5th SIGdial Workshop on Discourse and Dialogue (SIGdial).",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "Predicting student emotions in computer-human tutoring dialogues",
"authors": [
{
"first": "D",
"middle": [],
"last": "Litman",
"suffix": ""
},
{
"first": "K",
"middle": [],
"last": "Forbes-Riley",
"suffix": ""
}
],
"year": 2004,
"venue": "Proceedings of the 42nd Annual Meeting of the Association for Compuational Linguistics (ACL)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Litman, D. and K. Forbes-Riley. 2004b. Predicting student emotions in computer-human tutoring di- alogues. In Proceedings of the 42nd Annual Meet- ing of the Association for Compuational Linguis- tics (ACL). Barcelone, Spain.",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "A model of textual affect sensing using real-world knowledge",
"authors": [
{
"first": "Hugo",
"middle": [],
"last": "Liu",
"suffix": ""
},
{
"first": "Henry",
"middle": [],
"last": "Lieberman",
"suffix": ""
},
{
"first": "Ted",
"middle": [],
"last": "Selker",
"suffix": ""
}
],
"year": 2003,
"venue": "IUI '03: Proceedings of the 8th international conference on Intelligent user interfaces",
"volume": "",
"issue": "",
"pages": "125--132",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Liu, Hugo, Henry Lieberman, and Ted Selker. 2003. A model of textual affect sensing using real-world knowledge. In IUI '03: Proceedings of the 8th international conference on Intelligent user inter- faces. ACM Press, Miami, Florida, USA, pages 125-132.",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "Thumbs up? Sentiment classification using machine learning techniques",
"authors": [
{
"first": "Bo",
"middle": [],
"last": "Pang",
"suffix": ""
},
{
"first": "Lillian",
"middle": [],
"last": "Lee",
"suffix": ""
},
{
"first": "Shivakumar",
"middle": [],
"last": "Vaithyanathan",
"suffix": ""
}
],
"year": 2002,
"venue": "Proceedings of the 2002 Conference on Empirical Methods in Natural Language Processing (EMNLP)",
"volume": "",
"issue": "",
"pages": "79--86",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Pang, Bo, Lillian Lee, and Shivakumar Vaithyanathan. 2002. Thumbs up? Sentiment classification using machine learning techniques. In Proceedings of the 2002 Conference on Em- pirical Methods in Natural Language Processing (EMNLP). Philadelphia, Pennsylvania, pages 79-86.",
"links": null
},
"BIBREF17": {
"ref_id": "b17",
"title": "A brief introduction to boosting",
"authors": [
{
"first": "R",
"middle": [
"E"
],
"last": "Schapire",
"suffix": ""
}
],
"year": 1999,
"venue": "Proceedings of IJCAI",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Schapire, R.E. 1999. A brief introduction to boost- ing. In Proceedings of IJCAI.",
"links": null
},
"BIBREF18": {
"ref_id": "b18",
"title": "Improved boosting algorithms using confidence-rated predictions",
"authors": [
{
"first": "R",
"middle": [
"E"
],
"last": "Schapire",
"suffix": ""
},
{
"first": "Y",
"middle": [],
"last": "Singer",
"suffix": ""
}
],
"year": 1999,
"venue": "Machine Learning",
"volume": "37",
"issue": "3",
"pages": "297--336",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Schapire, R.E. and Y. Singer. 1999. Improved boosting algorithms using confidence-rated pre- dictions. Machine Learning 37(3):297-336.",
"links": null
},
"BIBREF19": {
"ref_id": "b19",
"title": "Smokey: Automatic recognition of hostile messages",
"authors": [
{
"first": "Ellen",
"middle": [],
"last": "Spertus",
"suffix": ""
}
],
"year": 1997,
"venue": "Proc. of Innovative Applications of Artificial Intelligence",
"volume": "",
"issue": "",
"pages": "1058--1065",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Spertus, Ellen. 1997. Smokey: Automatic recogni- tion of hostile messages. In In Proc. of Innova- tive Applications of Artificial Intelligence. pages 1058-1065.",
"links": null
},
"BIBREF20": {
"ref_id": "b20",
"title": "Measuring praise and criticism: Inference of semantic orientation from association",
"authors": [
{
"first": "P",
"middle": [],
"last": "Turney",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Littman",
"suffix": ""
}
],
"year": 2003,
"venue": "ACM Transactions on Information Systems",
"volume": "21",
"issue": "4",
"pages": "315--346",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Turney, P. and M. Littman. 2003. Measuring praise and criticism: Inference of semantic orientation from association. ACM Transactions on Informa- tion Systems 21(4):315-346.",
"links": null
},
"BIBREF21": {
"ref_id": "b21",
"title": "Recognizing contextual polarity in phraselevel sentiment analysis",
"authors": [
{
"first": "Theresa",
"middle": [],
"last": "Wilson",
"suffix": ""
},
{
"first": "Janyce",
"middle": [],
"last": "Wiebe",
"suffix": ""
},
{
"first": "Paul",
"middle": [],
"last": "Hoffmann",
"suffix": ""
}
],
"year": 2005,
"venue": "HLT '05: Proceedings of the conference on Human Language Technology and Empirical Methods in Natural Language Processing",
"volume": "",
"issue": "",
"pages": "347--354",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Wilson, Theresa, Janyce Wiebe, and Paul Hoffmann. 2005. Recognizing contextual polarity in phrase- level sentiment analysis. In HLT '05: Proceed- ings of the conference on Human Language Tech- nology and Empirical Methods in Natural Lan- guage Processing. Association for Computational Linguistics, Morristown, NJ, USA, pages 347- 354.",
"links": null
},
"BIBREF22": {
"ref_id": "b22",
"title": "Towards answering opinion questions: Separating facts from opinions and identifying the polarity of opinion sentences",
"authors": [
{
"first": "Hong",
"middle": [],
"last": "Yu",
"suffix": ""
},
{
"first": "Vasileios",
"middle": [],
"last": "Hatzivassiloglou",
"suffix": ""
}
],
"year": 2003,
"venue": "Proceedings of the Conference on Empirical Methods in Natural Language Processing",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Yu, Hong and Vasileios Hatzivassiloglou. 2003. To- wards answering opinion questions: Separating facts from opinions and identifying the polarity of opinion sentences. In Proceedings of the Confer- ence on Empirical Methods in Natural Language Processing (EMNLP).",
"links": null
}
},
"ref_entries": {
"FIGREF0": {
"num": null,
"type_str": "figure",
"text": "Email samples: 1) emotional; 2) neutral",
"uris": null
},
"FIGREF1": {
"num": null,
"type_str": "figure",
"text": "Precision/Recall curves for different experiments. Large black circles indicate the operating point with best F-Measure",
"uris": null
},
"TABREF1": {
"text": "",
"type_str": "table",
"num": null,
"html": null,
"content": "<table><tr><td colspan=\"3\">shows that emotional emails are about 12-13% of</td></tr><tr><td colspan=\"2\">the total population.</td><td/></tr><tr><td>Set</td><td colspan=\"2\">Number of examples Critical Emails</td></tr><tr><td>Training</td><td>620</td><td>12%</td></tr><tr><td>Test</td><td>457</td><td>13%</td></tr></table>"
},
"TABREF2": {
"text": "",
"type_str": "table",
"num": null,
"html": null,
"content": "<table/>"
},
"TABREF4": {
"text": "",
"type_str": "table",
"num": null,
"html": null,
"content": "<table><tr><td>: Recall and precision corresponding to best F-</td></tr><tr><td>measure for different classifier models</td></tr></table>"
}
}
}
} |