README File for the GIGAWORD CHINESE TEXT CORPUS ================================================ Second Edition ============== INTRODUCTION ------------ The Gigaword Chinese Corpus is a comprehensive archive of newswire text data that has been acquired over several years by the Linguistic Data Consortium (LDC), at the University of Pennsylvania. This is the second edition of the Gigaword Chinese Corpus. This edition includes all of the contents in the first edition of the Chinese Gigaword corpus (LDC2003T09) as well as new data collected after the publication of the first edition. Also, a limited number of articles from a new newspaper source (Zaobao) have been added in this edition. The three distinct international sources of Chinese newswire included in this edition are the following: - Central News Agency, Taiwan (cna_cmn) - Xinhua News Agency (xin_cmn) - Zaobao Newspaper (zbn_cmn) The seven-letter codes in the parentheses above include the three-character source name abbreviations and the three-character language code ("cmn") separated by an underscore ("_") character. The three-letter language code conforms to LDC's new internal convention based on the new ISO 639-3 standard. In the first edition of the Chinese Gigaword corpus and other previous LDC corpora, a simpler three-character-code scheme was used to identify both the source and the language. The new convention allows us to distinguish data sets by source and language more naturally when a single newswire provider distributes data in multiple languages. The following table shows the correspondence between the old codes and the new codes. new old ------------------ cna_cmn cna xin_cmn xin The new seven-letter codes are used in both the directory names where the data files are found, and in the prefix that appears at the beginning of every data file name. Some of the Xinhua content in this collection has been published previously by the LDC in other, older corpora, particularly Mandarin Chinese News Text (LDC95T13), TREC Mandarin (LDC2000T52), and the various TDT Multi-language Text corpora. WHAT'S NEW IN THE SECOND EDITION -------------------------------- New documents (Xinhua from October 2002 through December 2004, and CNA from January 2003 through December 2004) have been added. A new newspaper source (Lianhe Zaobao) has been added. CHARACTER ENCODING ------------------ The original data archives received by the LDC from Xinhua and Zaobao were encoded in GB-2312, whereas those from CNA were encoded in Big-5. To avoid the problems and confusion that could result from differences in character-set specifications, all text files in this corpus have been converted to UTF-8 character encoding. Researchers who have concerns about the comparability and compatibility of text data from GB and Big-5 sources should consult The Unicode Standard, version 3.0 (published by the Unicode Consortium, http://www.unicode.org), paying special attention to Chapter 10, "East Asian Scripts", and Appendix A, "Han Unification History". Owing to the use of UTF-8, the SGML tagging within each file (described in detail in the next section) shows up as lines of single-byte-per-character (ASCII) text, whereas lines of actual text data, including article headlines and datelines, contain a mixture of single-byte and multi-byte characters. Both Big-5 and GB are designed to support ASCII single-byte character data as well as 2-byte Chinese characters; in addition, each of these coding standards has a section of the 2-byte character space devoted to "full-width" renderings of the printable ASCII characters. For example, the digits 0-9 can be presented as either single-byte ASCII codes or as 2-byte full-width codes, as shown in the following table: Digit ASCII GB 2-byte Big-5 2-byte Character byte code-point code-point -------------------------------------------------- 0 0x30 0xA3C0 0xA2AF 1 0x31 0xA3C1 0xA2B0 2 0x32 0xA3C2 0xA2B1 3 0x33 0xA3C3 0xA2B2 4 0x34 0xA3C4 0xA2B3 5 0x35 0xA3C5 0xA2B4 6 0x36 0xA3C6 0xA2B5 7 0x37 0xA3C7 0xA2B6 8 0x38 0xA3C8 0xA2B7 9 0x39 0xA3C9 0xA2B8 and similarly for the upper- and lower-case alphabet characters, brackets, quotation marks and punctuation. We found that both archives showed evidence of somewhat free variation between single- and two-byte forms when presenting alphanumerics, etc, within the text data. Although the Unicode Standard provides an analogous portion of its code table to these full-width characters, we decided instead to eliminate this form of variation in the data: wherever the original data contained 2-byte versions of characters having exact correlates in the single-byte ASCII table, we replaced the 2-byte character with the single-byte ASCII equivalent. As a result, many lines of text data contain a mix of multi-byte Chinese and single-byte ASCII content. Of course, since all the data is now presented in UTF-8 encoding, this mixture is a natural property of the data, which any UTF-8-aware process will handle without difficulty. We also found that both archives used a handful of "accented" alphabetics and other special characters common to European character sets. When converted to UTF8, these characters assume their "normal" places in the Unicode table -- e.g. the "raised circle", used as a "degrees" mark in temperatures or latitude/longitude coordinates, can be found in the Xinhua data rendered as U00B0 (which in UTF8 form comes out as the two-byte sequence 0xC2 0xB0). Apart from these rare cases, all characters in the text are either single-byte ASCII or multi-byte Chinese. DATA FORMAT AND SGML MARKUP --------------------------- Each data file name consists of the 7-letter prefix (e.g., xin_cmn) and an underscore character ('_') followed by a 6-digit date (representing the year and month during which the file contents were generated by the respective news source), followed by a ".gz" file extension, indicating that the file contents have been compressed using the GNU "gzip" compression utility (RFC 1952). So, each file contains all the usable data received by LDC for the given month from the given news source. All text data are presented in SGML form, using a very simple, minimal markup structure. The file "gigaword_c.dtd" in the "docs" directory provides the formal "Document Type Declaration" for parsing the SGML content. The corpus has been fully validated by a standard SGML parser utility (nsgmls), using this DTD file. The markup structure, common to all data files, can be summarized as follows: The Headline Element is Optional -- not all DOCs have one The Dateline Element is Optional -- not all DOCs have one

Paragraph tags are only used if the 'type' attribute of the DOC happens to be "story" -- more on the 'type' attribute below...

Note that all data files use the UNIX-standard "\n" form of line termination, and text lines are generally wrapped to a width of 40 characters or less.

For every "opening" tag (DOC, HEADLINE, DATELINE, TEXT, P), there is a corresponding "closing" tag -- always. The attribute values in the DOC tag are always presented within double-quotes; the "id=" attribute of DOC consists of the 7-letter source and language ID (in CAPS), an underscore, an 8-digit date string representing the date of the story (YYYYMMDD), a period, and a 4-digit sequence number starting at "0001" for each date (e.g. "XIN_CMN_199501.0001"); in this way, every DOC in the corpus is uniquely identifiable by the id string. Every SGML tag is presented alone on one line, separate from other tags, and from the text content (so a simple process like the UNIX "grep -v '<'" will eliminate all tags, and retain all the text content). The structure shown above represents some notable differences relative to the markup strategy employed in previous LDC text corpora; these are intended to facilitate bulk processing of the present corpus. The major differences are: - Earlier corpora usually organized the data as one file per day, or limited the average file size to one megabyte (MB). Typical compressed file sizes in the current corpus range from about 500 KB (2003 CNA data) to about 10 MB (2001-2 CNA data); this equates to a range of about 1.5 to 27 MB when the data are uncompressed. In general, these files are not intended for use with interactive text editors or word processing software (though many such programs are likely to work reasonably well with these files). Rather, it's expected that the files will be used as input to programs that are geared to dealing with data in such quantities, for filtering, conditioning, indexing, statistical summary, etc. (The LDC can provide open source software, mostly written in Perl, for extracting DOCs from such data files, using the "id" string or other search criteria for story selection; see http://www.ldc.upenn.edu/Using/ .) - Earlier corpora tended to use different markup outlines (different tag sets) depending on the source of the data, because different sources came to us with different structural properties, and we had chosen to preserve these as much as possible (even though many elements of the delivered structure may have been meaningless for research use). The present corpus uses only the information structure that is common to all sources and serves a clear function: headline, dateline, and core news content (usually containing paragraphs). The "dateline" is a brief string typically found at the beginning of the first paragraph in each news story, giving the location the report is coming from, and sometimes the news service and/or date; since this content is not part of the initial sentence, we separate it from the first paragraph (this was not done in previous corpora). - Earlier corpora tended to include "custom" SGML entity references, which were intended to preserve things like special punctuation or typesetting instructions (e.g. "&QL;", "&UR;", "&MD;", etc). The present corpus uses only three SGML entity reference: - ``&'' represents the literal ampersand "&" character - ``<'' represents the literal open-angle bracket "<" - ``>'' represents the literal close-angle bracket ">" All other specialized control characters have been filtered out. - In earlier corpora, newswire data were presented as streams of undifferentiated "DOC" units; depending on the source and corpus, varying amounts of quality checking and filtering were done to eliminate noisy or unsuitable content (e.g. test messages). For this release, all sources have received a uniform treatment in terms of quality control, and we have applied a rudimentary (and _approximate_) categorization of DOC units into four distinct "types". The classification is indicated by the `` type="string" '' attribute that is included in each opening ``DOC'' tag. The four types are: * story : This is by far the most frequent type, and it represents the most typical newswire item: a coherent report on a particular topic or event, consisting of paragraphs and full sentences. As indicated above, the paragraph tag "

" is found only in DOCs of this type; in the other types described below, the text content is rendered with no additional tags or special characters -- just lines of ASCII tokens separated by whitespace. * multi : This type of DOC contains a series of unrelated "blurbs", each of which briefly describes a particular topic or event; this is typically applied to DOCs that contain "summaries of todays news", "news briefs in ... (some general area like finance or sports)", and so on. Each paragraph-like blurb by itself is coherent, but it does not bear any necessary relation of topicality or continuity relative to it neighbors. * advis : (short for "advisory") These are DOCs which the news service addresses to news editors -- they are not intended for publication to the "end users" (the populations who read the news); as a result, DOCs of this type tend to contain obscure abbreviations and phrases, which are familiar to news editors, but may be meaningless to the general public. We also find a lot of formulaic, repetitive content in DOCs of this type (contact phone numbers, etc). * other : This represents DOCs that clearly do not fall into any of the above types -- in general, items of this type are intended for broad circulation (they are not advisories), they may be topically coherent (unlike "multi" type DOCS), and they typically do not contain paragraphs or sentences (they aren't really "stories"); these are things like lists of sports scores, stock prices, temperatures around the world, and so on. The general strategy for categorizing DOCs into these four classes was, for each source, to discover the most common and frequent clues in the text stream that correlated with the three "non-story" types, and to apply the appropriate label for the ``type=...'' attribute whenever the DOC displayed one of these specific clues. When none of the known clues was in evidence, the DOC was classified as a "story". This means that the most frequent classification error will tend to be the use of `` type="story" '' on DOCs that are actually some other type. But the number of such errors should be fairly small, compared to the number of "non-story" DOCs that are correctly tagged as such. Note that the markup was applied algorithmically, using logic that was based on less-than-complete knowledge of the data. For the most part, the HEADLINE, DATELINE and TEXT tags have their intended content; but due to the inherent variability (and the inevitable source errors) in the data, users may find occasional mishaps where the headline and/or dateline were not successfully identified (hence show up within TEXT), or where an initial sentence or paragraph has been mistakenly tagged as the headline or dateline. DATA QUANTITIES --------------- The "docs" directory contains a set of plain-text tables (datastats_*) that describe the quantities of data by source and month (i.e. by file), broken down according to the four "type" categories. The overall totals for each source are summarized below. Note that the "Totl-MB" numbers show the amount of data you get when the files are uncompressed (i.e. nearly 4 gigabytes, total); the "Gzip-MB" column shows totals for compressed file sizes as stored on the DVD-ROM; the "K-wrds" numbers are actually the number of Chinese characters (there is no notion of "space separated word tokens" in Chinese): Source #Files Gzip-MB Totl-MB K-wrds #DOCs CNA_CMN 168 1052 2748 792195 1769953 XIN_CMN 168 642 1603 471110 992261 ZBN_CMN 10 39 93 28066 41418 TOTAL 346 1733 4444 1291371 2803632 The following tables present "K-wrds" (i.e. thousands of Chinese characters) and "#DOCS" broken down by source and DOC type: #DOCs K-wrds type="advis": CNA_CMN 8160 751 XIN_CMN 6553 711 ZBN_CMN 0 0 TOTAL 14713 1462 type="multi": CNA_CMN 30552 23429 XIN_CMN 11329 7516 ZBN_CMN 55 41 TOTAL 41936 30986 type="other": CNA_CMN 100758 40258 XIN_CMN 31255 9999 ZBN_CMN 279 130 TOTAL 132292 50387 type="story": CNA_CMN 1630483 727748 XIN_CMN 943132 452878 ZBN_CMN 41084 27898 TOTAL 2614691 1208524 GENERAL PROPERTIES OF THE DATA ------------------------------ All of the data sets have been produced from bulk archives that were delivered to the LDC via internet transfer. As a result, we avoided many of the problems that commonly afflict newswire data that has been transmitted over modems. Still, both archives contained noticeable amounts of "noise" (unusable characters, null bytes, etc) which had to be filtered out for research use. Two of the corpus authors at the LDC, Ke Chen and Junbo Kong, are native speakers of Mandarin Chinese, and did extensive diagnosis to identify and eliminate unsuitable content in the original archival data. To some extent, this is an open-ended problem, and there may be kinds of error conditions that have gone unnoticed or untreated -- this is true of any large text collection -- but we have striven to assure that the characters presented in all files are in fact valid and displayable, and that the markup is fully SGML compliant. README file written by David Graff and Ke Chen, January 2003 Updated for the Second Edition by Junbo Kong and Kazuaki Maeda, June 2005. Linguistic Data Consortium