Part Two: Using Xaira to explore corpora

1
Part TwoUsing Xaira to explore corpora

• Richard Xiao
• z.xiao_at_lancaster.ac.uk

2
Outline of the talk

• Concordance
• Wordlist
• Keywords (No)
• Output formats
• Manipulating results
• Collocation/colligation
• Distribution analysis
• Live demonstration
• Tips for keeping away from bugs
• Multilingual dimension
• Xaira FAQs

3
Concordance

• Word query ( )
• Search for a word
• Phrase/Quick query ( or
  )
• Searching for a word or phrase
• Addkey query ( )
• POS or lemma search
• Pattern query ( )
• Regular Expression search
• XML query ( )
• Search for XML markup
• CQL/XQL query ( )
• Searching using XML-based Corpus Query Language
• Query builder ( )
• A powerful combination of all query types

4
Wordlist

• In Client gtgt Word query (up to 100,000 lexicon
  entries) sorting alphabetically, by frequency,
  or the number of forms
• In Xaira Indexer Tools gtgt Tools gtgt Indexer gtgt
  Options gtgt Create frequency table

5
Keyword?

• Sadly, no
• Use WordSmith instead
• WordSmith version 4.0 fully supports Unicode

6
Output formats

• Page mode vs. Line mode (KWIC)
• Plain text vs. XML text
• Scope of context
• Alignment (left, right, top, bottom)
• Reference (on the status bar)

7
Manipulating results

• Edit query (to save time for related queries)
• Bibliographical data
• Sort KWIC concordances
• Select/block select/copy concordances
• Right click on a concordance
• Thin/edit concordances
• Random sampling
• Save queries and export them in XML
• Print results

8
Collocation/colligation ( )

• Statistical measure (MI or Z)
• Window span
• Minimum frequency
• Minimum MI/Z score
• Top N collocates
• Computing collocation statistics for individual
  words
• Applying selected lemmata
• Colligation (Addkey tags)

9
Distribution analysis ( )

• Defining partition (subcorpora)
• (Texts gtgt Column control to select XML tags)
• Texts gtgt Define partition (3 ways)
• Based on selected class, values in a column, or
  solutions to a query
• Texts gtgt Open partition
• Tabulation (text class, words, hits, , etc)
• Normalised frequencies for subcorpora
• Sorting tabulated data
• Graphic presentation (pie/bar chart)
• Save distribution data in various forms
• Copy pie/bar chart

10
Additional features of Xaira

• Annotating concordances (making notes)
• Copying query text or notes
• User-defined stylesheet
• Colour book (e.g. different colours for different
  POS categories)
• Remote access over a network
• Platform-independent

11
Xaira live demonstration

• Here we go
• slides to follow

12
Tips for keeping away from bugs

• In the Line mode, a maximum of 1,524 concordances
  are displayed
• See the rest in the Page mode
• In Query builder, joining query nodes in the
  horizontal direction (OR) and then in the
  vertical direction (AND) may produce unreliable
  counts when the Link type is specified as
  One-way or Two-way
• Only define Link type as Next or Not next
• If thousands of hits are downloaded and dozens of
  them are deleted by reverse selection in
  thinning, the system may crash
• If concordances have been sorted/edited, a saved
  query may not be opened again
• Save the edited concordances as an XML list using
  Query Listing in the menu or pressing on
  the toolbar

13
Truly multilingual - Chinese
14
Truly multilingual - Bengali
15
Truly multilingual - Hindi
16
Truly multilingual - Punjabi
17
Truly multilingual - Urdu
18
Xaira FAQs

• Is Xaira free and where can I get it?
• Yes, it is absolutely free. You can get a copy
  (binary for Windows, and source codes for
  compilation on the Unix/Linux/Mac system) at the
  SourceForce website. The latest release is 116.
  http//sourceforge.net/project/showfiles.php?group
  _id130289
• Where can I get more documentation?
• In addition to the built-in help file, more
  documentation is available at the Xaira site
  http//www.oucs.ox.ac.uk/rts/xaira/
• Where can I get technical help?
• You can sign up for the Xaira Preview List to get
  help http//www.tei-c.org.uk/tei-bin/betatest
• For a critical review, see http//www.lancs.ac.uk/
  postgrad/xiaoz/papers/xaira_review.pdf