{ "paper_id": "C90-1005", "header": { "generated_with": "S2ORC 1.0.0", "date_generated": "2023-01-19T12:36:01.500996Z" }, "title": "Tagging for Learning: Collecting Thematic Relations from Corpus", "authors": [ { "first": "Uri", "middle": [], "last": "Zernik", "suffix": "", "affiliation": {}, "email": "" }, { "first": "Paul", "middle": [], "last": "Jacobs", "suffix": "", "affiliation": {}, "email": "" } ], "year": "", "venue": null, "identifiers": {}, "abstract": "Recent work in text analysis has suggested that data on words that frequently occur together reveal important information about text content. Co-occurrence relations can serve two main purposes in language processing. First, the statistics of co-occurrence have been shown to produce accurate results in syntactic analysis. Second, the way that words appear together can help in assigning thematic roles in semantic interpretation. This paper discusses a method for collecting co-occurrence data, ~quiring lexical relations from the data, and applying these relations to semantic analysis.", "pdf_parse": { "paper_id": "C90-1005", "_pdf_hash": "", "abstract": [ { "text": "Recent work in text analysis has suggested that data on words that frequently occur together reveal important information about text content. Co-occurrence relations can serve two main purposes in language processing. First, the statistics of co-occurrence have been shown to produce accurate results in syntactic analysis. Second, the way that words appear together can help in assigning thematic roles in semantic interpretation. This paper discusses a method for collecting co-occurrence data, ~quiring lexical relations from the data, and applying these relations to semantic analysis.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Abstract", "sec_num": null } ], "body_text": [ { "text": "Two text processing problems rely heavily on cooccurrence patterns-the way that words appear together, possibly idiosyncraticly. First, statistically weighted co-occurrence information can assist in the \"bracketing\" of noun groups, which can otherwise lead to a eombinatoric explosion of parse trees [1] . Second, co-occurrence relations can provide evidence of semantic information for thematic-role assignment, an important task that is otherwise fraught with inaccuracy.", "cite_spans": [ { "start": 300, "end": 303, "text": "[1]", "ref_id": "BIBREF0" } ], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "Only co-occurrence patterns collected over a corpus can help to determine which is .object and which is recipient in PAID DIVIDEND (IS SECURE) vs. PAID", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": ". A sufficiently rich lexicon would include the semantic preferences for distinguishing these thematic roles, but such a lexicon does not yet exist. Co-occurrence patterns are a means of probing a global corpus for clues that help resolve ambiguity at the local sentence level. Patterns such as PAID TO SHAREHOLDERS and PAID THEM THE DIVIDEND are detected in the corpus at large. Through these latter examples, in which the distinction between recipient and object relative to the dative verb PAY is made explicit, the former cases in which tile relation is implicit can be resolved.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "SHAREHOLDERS (ARE SATISFIED)", "sec_num": null }, { "text": "In contrast to previous work which addressed the identification of surface relations, i.e., SVO triples [2] , in our work we address the acquisition of semantic relations, focussing at the assigment of thematic roles. This task (i.e. tagging for acquisition) requires high reliability and so it relies less on statistical properties and more on deterministic local marking.", "cite_spans": [ { "start": 104, "end": 107, "text": "[2]", "ref_id": "BIBREF1" } ], "ref_spans": [], "eq_spans": [], "section": "SHAREHOLDERS (ARE SATISFIED)", "sec_num": null }, { "text": "In this paper we discuss a technique for parsing and semanticly analyzing complex sentences with the aid of co-occurrence relations, and show how these relations are acquired from tagged corpus.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "SHAREHOLDERS (ARE SATISFIED)", "sec_num": null }, { "text": "Consider, for example, the sentence below, taken from the Dow-Jones newswire:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "The Phenomenon", "sec_num": "1.1" }, { "text": "For this sentence, which is not exotic or unusual in its complexity, there are 24 non-trivial different parse trees. Human readers, in contrast to most programs, can quickly identify groups of words that \"hang together\" such as COMPANY PAID A DIVI-DEND, STOCK DIVIDEND, and CASH DIVIDEND, and use these clusters to understand the sentence unambiguously. Moreover, a human reader can easily recognize SHAREHOLDERS as recipient and DIV-IDEND as the object of PAY. Along these lines, our program develops the capability to identify such patterns by training on a large corpus of examples.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "THE LARGEST CO~iPANY ON THE LIST, WHICH LAST PAID SHAREHOLDERS IN JANUARY, SAID THE 5 PC STOCK DIVIDEND WOULD BE PAYABLE FOLLOWING THE PAYMENT OF THE CASH DIVIDEND. (DJ, October 27, 1988)", "sec_num": null }, { "text": "The training corpus, from which our lexical information is extracted, consists of more than ten rail-lion words from the Dow Jones newswire (10 months worth of stories). For the root PAY, for instance, we collected more than 6000 examples, 20 of which are given below.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "The Training Corpus", "sec_num": "1.2" }, { "text": "To exploit this data, a system must transform common patterns into operational templates, encoding a core relation between the words. The sections that follow describe the evolution and implementation of this acquisition technique. CAR is selected over ROAD as the anaphor of IT, since CAR BRAKE is a stronger collocation than ROAD BRAKE. Interestingly, this idea complements Wilks' preference semantics [8] , in which preference is based on a semantic hierarchy. In Dagan's method, preferences are based on word patterns acquired from corpus.", "cite_spans": [ { "start": 376, "end": 382, "text": "Wilks'", "ref_id": null }, { "start": 404, "end": 407, "text": "[8]", "ref_id": "BIBREF7" } ], "ref_spans": [], "eq_spans": [], "section": "The Training Corpus", "sec_num": "1.2" }, { "text": "Our work further emphasizes globM-sentence connections. An example that highlights the use of cooccurrence is given on the next page.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "The Training Corpus", "sec_num": "1.2" }, { "text": "What is the attachment of THAT? THAT could potentially attach to almost any preceding word, e.g., FEDERAL THAT, BOARD THAT, CONVERSION THAT, SAID THAT, etc. The affinity of the word pair SAY THAT (although it does not appear in this sentence as a collocation) supports the appropriate attachment.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "THE CHAIRMAN AND CHIEF EXECUTIVE OF FRANKL-IN FIRST FEDERAL SAVINGS ~ LOAN ASSOCIAT-ION OF WILKES-BARRE, [SAID] FRANKLIN FIRST FEDERAL'S PLAN OF CONVERSION HAD BEEN APPROVED BY THE FEDERAL HOME LOAN BANK BOARD [AND THAT] THE OFFERING OF COMMON SHARES IN FRANKLIN FIRST FINANCIAL CORP. HAD BEEN APPROVED BY THE BANK BOARD AND BY THE SEC. (D J, 07-25-88).", "sec_num": null }, { "text": "Furthermore, co-occurrence relations support thematic-role assignment. This is important for our ultimate objective of producing more accurate conceptual information from news stories [5] . The text below illustrates one type of problem in role assignment:", "cite_spans": [ { "start": 184, "end": 187, "text": "[5]", "ref_id": "BIBREF4" } ], "ref_spans": [], "eq_spans": [], "section": "THE CHAIRMAN AND CHIEF EXECUTIVE OF FRANKL-IN FIRST FEDERAL SAVINGS ~ LOAN ASSOCIAT-ION OF WILKES-BARRE, [SAID] FRANKLIN FIRST FEDERAL'S PLAN OF CONVERSION HAD BEEN APPROVED BY THE FEDERAL HOME LOAN BANK BOARD [AND THAT] THE OFFERING OF COMMON SHARES IN FRANKLIN FIRST FINANCIAL CORP. HAD BEEN APPROVED BY THE BANK BOARD AND BY THE SEC. (D J, 07-25-88).", "sec_num": null }, { "text": "Who paid what to whom and when? Cooccurrence-based analysis generates lexical relations such as subj-verb, verb-obj, and verb-obj2, relations which are further mapped into appropriate thematic and semantic roles. The program thus determines that COMPANY is the payer of PAID, SHAREHOLD-ERS the payee, and DIVIDEND the payment.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "THE LARGEST COMPANY ON THE LIST, WHICH LAST PAID SHAREHOLDERS IN JANUARY, SAID THE 5 PC STOCK DIVIDEND WOULD BE PAYABLE FOLLOWING THE PAYMENT OF THE CASH DIVIDEND. (D J, October 27, 1988)", "sec_num": null }, { "text": "Lexical Representation", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "3", "sec_num": null }, { "text": "An acquired lexical structure called a Thematic Relations ( Figure 2 ) facilitates this analysis. For a pair of content words, a relation provides (1) a strength of association (or \"mutual affinity\"), and (2) a structure type. This table is acquired from corpus by a tagger based on morphology and local syntax.", "cite_spans": [], "ref_spans": [ { "start": 60, "end": 68, "text": "Figure 2", "ref_id": null } ], "eq_spans": [], "section": "3", "sec_num": null }, { "text": "Extracting Co-occurrence Relations from Corpus", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "4", "sec_num": null }, { "text": "The algorithm operates in three steps: (1) tag the corpus for morphology and part of speech, (2) collect collocations using relative frequency, and (3) use tagging to determine lexical relations within collocations.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "4", "sec_num": null }, { "text": "Since the corpus size is about 10-million words, a fullfledged global sentence parsing is prohibitively expensive, and tagging must be carried out by localist methods, i.e., by means of morphology and local syntactic markers. There are three degrees of difficulty of cases to be tagged. ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Part-of-speech Tagging", "sec_num": "4.1" }, { "text": "Some cases prove even more difficult and cannot be resolved by localist methods. Consider the following two examples.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Problematic Cases", "sec_num": "4.2" }, { "text": "\u2022 \"The company preferred stock PAID ...\" . In this clause, PAID, could be either an adjective or a verb (see \"the horse raced past the barn\"). Indeed, this clause could probably be determined by a global parse, however, this would be too expensive computationally.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Problematic Cases", "sec_num": "4.2" }, { "text": "\u2022 \"CONVINCING MANAGEMENT proved tough\" is even harder since it presents a Necker cube situation (i. e. changing the interpretation of either word seems immediately to change the interpretation of the pair). Is it an adjective-noun or is it a verb-noun pair? In general, the analysis of such pairs requires deeper understanding of word relationships. Consider another example: The incorrect resolution of such cases, which unfortunately are pervasive in the corpus, impinges on two objectives: performance and learning.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Problematic Cases", "sec_num": "4.2" }, { "text": "LATER", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Problematic Cases", "sec_num": "4.2" }, { "text": "In order to perform text analysis, in the first case one must determine whether management was convinced, or the management convinced some second party; in the second case, one must determine the subject of the main verb of the sentence, i.e., which is the ,subject of DIMINISHED? Many applications require an unambiguous result. Thus a call must be made one way or another. Statistical means might make that call slightly more judiciuos on the average.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Problematic Cases", "sec_num": "4.2" }, { "text": "However, when tagging is used for learning of thematic roles, inappropriate resolution of such cases can drastically contaminate the final results by biasing it in a certain direction. Results are far more accurate when ambiguous cases are left out altogether.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Problematic Cases", "sec_num": "4.2" }, { "text": "Our tagger is based on a 7,000-root lexicon that facilitates accurate morphological analysis, and about I00 local-syntax rules. It produces tagging for about 60% of the content words in the corpus. Tagged output for a sample sentence is given below. ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Tagging for Learning", "sec_num": "4.3" }, { "text": "*period*///SP A 4-tuple in the sentence above is a word/root/affix/part-of-speech. As expected, many content words in this sentence cannot be unambiguously tagged, and are marked ?, i.e., undetermined. In particular, notice that PAID remains unresolved.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "PC///NN STOCK///?? DIVIDEND///NN WOULD/WII~ //AX BE///AX PAYABLE/PAY/ABLE/AD FOLLOWING/ FOLLOW/ING/?? THE///DT PAYMENT/PAY/HENT/NN OF///PP THE///DT CASH///NN DIVIDEND///NN", "sec_num": null }, { "text": "Fortunately, most PAY cases in the corpus are simpler and are appropriately tagged.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "PC///NN STOCK///?? DIVIDEND///NN WOULD/WII~ //AX BE///AX PAYABLE/PAY/ABLE/AD FOLLOWING/ FOLLOW/ING/?? THE///DT PAYMENT/PAY/HENT/NN OF///PP THE///DT CASH///NN DIVIDEND///NN", "sec_num": null }, { "text": ". ..", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "NUARYIIIDD", "sec_num": null }, { "text": "For purposes of thematic role acquisition the identification of passive and active voice is crucial. In the sample sentence above, PAID is appropriately tagged as a verb in the active voice (marked as VA).", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "NUARYIIIDD", "sec_num": null }, { "text": "Based on the tagging above (the root field), all collocations in the corpus are counted, and the following table is generated.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Collecting Collocations", "sec_num": "4.4" }, { "text": "This table is similar to Smadja's [7] , and it provides the position of collocative words relative to PAY, and the total count within 4 words in either direction.", "cite_spans": [ { "start": 34, "end": 37, "text": "[7]", "ref_id": "BIBREF6" } ], "ref_spans": [], "eq_spans": [], "section": "Collecting Collocations", "sec_num": "4.4" }, { "text": "Lexical relations are determined using the known functionality of the verb (see [9] ) and supporting examples. PAY is marked in the lexicon as a dative verb.", "cite_spans": [ { "start": 80, "end": 83, "text": "[9]", "ref_id": "BIBREF8" } ], "ref_spans": [], "eq_spans": [], "section": "Determining Lexical Relations", "sec_num": "4.5" }, { "text": "Consider 5 cases containing the pair PAY SHARE-HOLDER, from which the thematic relation is induced (VA stands for verb, active voice; VP for verb, passive voice; AD for adjective).", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Determining Lexical Relations", "sec_num": "4.5" }, { "text": "Exanlples (1), (4), and (5) support the hypothesis that StIAREHOLDER is an object2 (the recipient) of PAY.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Determining Lexical Relations", "sec_num": "4.5" }, { "text": "Based on a number of tagged sentences, the system determines that SHAREHOLDERS are recipients of PAY, while DIVIDENDS axe objects. This generalized lexical relation enables the semantic resolution of more difficult cases such as DIVIDEND PAYMENT and COMPANY PAID STOCK DIVIDEND.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Current Status and Conclusions", "sec_num": "5" }, { "text": "The implemented system using these techniques includes several elements: (1) morphology analysis currently produces accurate results for all the required cases; (2) tagging -produces results for only 60% of the required examples; more detailed rules could improve this figure to about 70%; (3) rule forming -currently works only with dative verbs such as PAY and SELL.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Current Status and Conclusions", "sec_num": "5" }, { "text": "A number of important pieces of recent research have highlighted the power of co-occurrence information in text. In the techniques described here, we have extended this research to use co-occurrence information for discriminating thematic roles. These techniques combine data acquisition from a tagged corpus with relation-driven language analysis to derive thematic knowledge from the text.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Current Status and Conclusions", "sec_num": "5" } ], "back_matter": [ { "text": " word -4 -3 -2 -1 0 +1 +2 +3 +4 total PRICE 5 14 438 38 0 17 12 32 12 558 COMPANY 47 53 71 26 0 2 6 1 161 367 DIVIDEND 37 42 36 121 0 11 1 14 25 287 RATE 6 5 16 109 0 14 112 16 3 ", "cite_spans": [], "ref_spans": [ { "start": 1, "end": 231, "text": "word -4 -3 -2 -1 0 +1 +2 +3 +4 total PRICE 5 14 438 38 0 17 12 32 12 558 COMPANY 47 53 71 26 0 2 6 1 161 367 DIVIDEND 37 42 36 121 0 11 1 14 25 287 RATE 6 5 16 109 0 14 112 16 3", "ref_id": null } ], "eq_spans": [], "section": "annex", "sec_num": null } ], "bib_entries": { "BIBREF0": { "ref_id": "b0", "title": "Hindie. Parsing, word associations, and predicateargument relations", "authors": [ { "first": "K", "middle": [], "last": "Church", "suffix": "" }, { "first": "W", "middle": [], "last": "Gale", "suffix": "" }, { "first": "P", "middle": [], "last": "Hanks", "suffix": "" }, { "first": "D", "middle": [], "last": "", "suffix": "" } ], "year": 1989, "venue": "Proceedings of the International Workshop on Parsing Technologies", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "K. Church, W. Gale, P. Hanks, and D. Hin- die. Parsing, word associations, and predicate- argument relations. In Proceedings of the In- ternational Workshop on Parsing Technologies, Carnegie Mellon University, 1989.", "links": null }, "BIBREF1": { "ref_id": "b1", "title": "Using statistics in lexical analysis", "authors": [ { "first": "K", "middle": [], "last": "Church", "suffix": "" }, { "first": "W", "middle": [], "last": "Gale", "suffix": "" }, { "first": "P", "middle": [], "last": "Hanks", "suffix": "" }, { "first": "D", "middle": [], "last": "Hindle", "suffix": "" } ], "year": 1990, "venue": "Lezical Acquisition: Exploiting On-Line Resources", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "K. Church, W. Gale, P. Hanks, and D. Hindle. Us- ing statistics in lexical analysis. In U. Zernik, ed- itor, Lezical Acquisition: Exploiting On-Line Re- sources. Lawrence Erlbanm Associates, Hillsdale, NJ, 1990.", "links": null }, "BIBREF2": { "ref_id": "b2", "title": "Using collocation in anaphora resolution", "authors": [ { "first": "I", "middle": [], "last": "Dagan", "suffix": "" } ], "year": 1989, "venue": "Computer Science Deptartment", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "I. Dagan. Using collocation in anaphora resolu- tion. Technical report, Technion, Computer Sci- ence Deptartment, Haifa, Israel, 1989.", "links": null }, "BIBREF3": { "ref_id": "b3", "title": "The Computational Analysis of English. Longman, London", "authors": [ { "first": "G", "middle": [], "last": "Leech", "suffix": "" }, { "first": "R", "middle": [], "last": "Garside", "suffix": "" }, { "first": "G", "middle": [], "last": "Sampson", "suffix": "" } ], "year": 1987, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "G. Leech R. Garside and G. Sampson. The Com- putational Analysis of English. Longman, Lon- don, Britain, 1987.", "links": null }, "BIBREF4": { "ref_id": "b4", "title": "Information extraction and text summarization using linguistic knowledge acquisition. Information Processing and Management", "authors": [ { "first": "Lisa", "middle": [ "F" ], "last": "Ran", "suffix": "" }, { "first": "Paul", "middle": [ "S" ], "last": "Jacobs", "suffix": "" }, { "first": "Uri", "middle": [], "last": "Zernik", "suffix": "" } ], "year": 1989, "venue": "", "volume": "25", "issue": "", "pages": "419--428", "other_ids": {}, "num": null, "urls": [], "raw_text": "Lisa F. Ran, Paul S. Jacobs, and Uri Zernik. Information extraction and text summarization using linguistic knowledge acquisition. Informa- tion Processing and Management, 25(4):419-428, 1989.", "links": null }, "BIBREF5": { "ref_id": "b5", "title": "Annotation manual for the pen treebank project", "authors": [ { "first": "B", "middle": [], "last": "Santorini", "suffix": "" } ], "year": 1990, "venue": "Computer and Information Science", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "B. Santorini. Annotation manual for the pen treebank project. Technical report, University of Pennsylvania, Computer and Information Sci- ence, Philadelphia, PA, 1990.", "links": null }, "BIBREF6": { "ref_id": "b6", "title": "Macrocoding the lexicon with cooccurrence knowledge", "authors": [ { "first": "F", "middle": [], "last": "Smadja", "suffix": "" } ], "year": 1989, "venue": "First International Le~:ical Acquisition Workshop", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "F. Smadja. Macrocoding the lexicon with co- occurrence knowledge. In U. Zernik, editor, First International Le~:ical Acquisition Workshop. 1989.", "links": null }, "BIBREF7": { "ref_id": "b7", "title": "A preferential, pattern-matching semantics for natural language understanding", "authors": [ { "first": "Y", "middle": [], "last": "Wilks", "suffix": "" } ], "year": 1975, "venue": "Artificial Intelligence", "volume": "6", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Y. Wilks. A preferential, pattern-matching se- mantics for natural language understanding. Ar- tificial Intelligence, 6, 1975.", "links": null }, "BIBREF8": { "ref_id": "b8", "title": "Lexical acquisition: Learning from corpus by capitalizing on lexieal categories", "authors": [ { "first": "U", "middle": [], "last": "Zernik", "suffix": "" } ], "year": 1989, "venue": "Proceedings of the Eleventh International Joint Conference on Artificial Intelligence", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "U. Zernik. Lexical acquisition: Learning from cor- pus by capitalizing on lexieal categories. In Pro- ceedings of the Eleventh International Joint Con- ference on Artificial Intelligence, Detroit, Michi- gan, 1989.", "links": null } }, "ref_entries": { "FIGREF0": { "uris": null, "num": null, "type_str": "figure", "text": "IN THE DAY BUYING INTEREST DIMINISHED . . . Again, it is difficult to tell whether INTEI~EST in BUYING diminished or the BUYING of IN-TERESTs diminished. Thus, local clues do not contribute towards the proper resolution of such cc'~3es." }, "FIGREF1": { "uris": null, "num": null, "type_str": "figure", "text": "THE///DT LIST///ml *co~ma*///SP WHICH///CC LAST///AD PAID/PAY/ED/? SHAREHO-LDEmS/SHARZHOLDER/S/NN IN///PP JANUARY///DD *comma*///CC SAID/SAY/ED/?? THE///DT 8//lAD" }, "TABREF1": { "num": null, "content": "
IT THE COMPANY LAST A 10 PC STOCK DIVIDEND TO BE JUNE 30. N INCOME DIVIDEND OF 1C A SHR AUG. 1S. THE COMPANY LAST UT THE SPECIAL DIVIDEND TO BE CT. 21. THE COMPANY LAST 10 PER SHARE SPECIAL DIVIDEND PER SHARE. THE DIVIDEND IS TED FOR A 5 PC STOCK DIVIDEND ERLY DIVIDEND OF 68.75 CENTS TERLY DIVIDEND OF 12 CENTS IS HE SPLIT AND THE DIVIDEND ARE 1.5 MILLION. THE DIVIDEND IS F THE COMPANY ON ANY DIVIDEND N THE UPCOMING FINAL DIVIDEND LDING ONE ADDITIONAL DIVIDEND OCT. I Figure h PAY Sentences in Corpus PAID A 7.5C DIVIDEND ON MAY 9. PAID AUG. 18. -0-; 2 09 PM EDT 07-28-88:\"? -O-GROW GROUP PAID IN FEBRUARY. -0-; 3 10 PM EDT 07-28-88: PAID A lOC SPECIAL DIVIDEND IN SEPTEMBER 1987 PAID FROM PROCEEDS OF THE SALE TO $6 A SHARE PAID A DIVIDEND OF 11 CENTS A SHARE ON J~Y 2 PAID TO STOCKHOLDERS ON JAN. 5, 1988. TOPP PAYABLE PAYABLE PAYABLE PAYABLE PAYABLE PAYABLE PAYMENT PAYMENT PAYMENT TO SHAREHOLDERS OF RECORD JULY 5. AUG. 12 TO HOLDERS OF RECORD JULY 15. 0.15 0 56 0 73 0 11 0 19 0 22 0 46 predicate:PAY predicate:PAY predicate:PAY predicate:PAY predicate:PAY predicate:PAY predicate:PAY subject:COMPANY object:DIVIDEND object2:SHAREHOLDER object:MILLION object:CASH object:*number* PC object:TOP RATES 1] have capitalized on a large collection of bi-grams plus statistically weighted grammar rules. In this method, statistical properties are ac-quired from a large training corpus which was tagged manually. Statistical methods have proved very effective, and attained a high level Figure 2: Statistics-Based Tagging: Taggers reported by [4; of accuracy [6].
", "type_str": "table", "text": ", IT HAS AGREED NOT TO D THAT IT INTENDS TO CONTINUE TIONS AND MODIFIYING DIVIDEND A PATTERN FOR THE FUTURE. PAY ANY FUTURE CASH DIVIDENDS, INCLUDING THE PAYING THEDIVIDEND. -0-; 11 08 AM EDT 07-22-PAYING A STOCK OF 60 CENTS FOR A TOTAL OF $1. PAID A SPECIAL DIVIDEND OF 8C LAST YEAR. WILL BE PAYED IN THE USUAL MAN AUG. 29 TO HOLDERS OF RECORD AUG. 12. SEPT. 14 TO HOLDERS OF RECORD AUG. 22 AUG. 18 TO HOLDERS OF RECORD AUG. 8. DATE ON OR AFrER AUG. 1, 1990, FOR TH OF 10.85 PENCE A SHARE. HEIGHTENING OVER A 12-MONTH PERIOD. DUE THURSDAY. Word Pairs Indicating Mutual Affinity and Thematic RolesSyntax-based Tagging: Local syntactic markers help to remove most cases of ambiguity. For example, was SAID (read: the word SAID preceded by was) can be unambiguously tagged a verb; the PAID shareholders, is an adjective; and the STOCK is definitely a noun.", "html": null } } } }