MAIQuery and TM Navtree

Download Report

Transcript MAIQuery and TM Navtree

Leveraging Your
Taxonomy to
Increase User
Productivity
MAIQuery and TM Navtree
Taxonomies aid site organization
Taxonomy provides:
 Framework for content
organization
 Hierarchical outline of your
content by subject categories
 Basis for effective browsing
2
Integrated taxonomy enhances
findability

Browsable categories of a directory

Smart search for term equivalents

Taxonomy terms (original or
modified) as labels

Navigation aids incorporate taxonomy
terms and relationships
3
Example Search: body growth
Complete database (60,000 + titles)
 Free text search
8

Free text search on titles
6

hits — some irrelevant
hits — limited recall
Search by taxonomy descriptor (AKA
subject term or category)
 470
hits
 100%
relevant
 100% recall
4
Increasing User Productivity

Items in an information collection
can be retrieved with better
precision (relevance) and better
recall by using a controlled
vocabulary to assign subject terms
(key words) to them
How do you connect your
users to the controlled
vocabulary?
5
Connecting Users
1.
Use the rulebase you’ve
developed for machine
aided indexing (MAIQuery)
2.
Use the controlled
vocabulary itself
(TM Navtree)
6
MAI’s talents
 MAI
(Machine Aided Indexer)
 helps
authors and editors
assign effective subject terms
 automates
the assignment of
subject terms to items in legacy
collections
7
Taxonomy terms on documents
help sort and organize the content
M.A.I. suggests the correct terms
from the taxonomy as descriptors
 M.A.I. rulebase recognizes term
equivalents

germs  Microorganisms
 vaccin*  Pharmaceutical drugs

Recognizing term equivalents
enables enhanced search
8
MAI’s “hidden talents”
 MAI
can also:
 Provide for the appropriate
preferred term when given a
word or phrase
 Return preferred terms for uses
of the word in different
contexts
9
More “hidden talents”
 MAIQuery
 Show
can:
related terms from the
thesaurus to broaden a search
 Show the rules and preferred
term’s scope notes to clarify
how the preferred term relates
to others in the thesaurus
10
Presenting: MAIQuery™
 Web
page presents a search box
that will use the MAI rulebase
 Can
be in addition to full text
search and advanced search
 User
enters a word or phrase in
the search box
 MAI searches the rulebase for
any occurrences of the word(s)
11
MAIQuery
12
the MAIQuery demo
 Uses
web pages and php coding:
 Passes
the search words to
“dosearch.php”
 dosearch.php passes the term to
MAI’s concept extractor
 MAI returns a list of suggested
terms from the controlled
vocabulary
13
Suggested terms
The term Music is suggested
by the rule for music*(1)
Click on the first (the preferred term) to see the term record; click on the second to see the MAI rule
The term Instrumental Music is suggested
by the rule for music*(1)
Click on the first (the preferred term) to see the term record; click on the second to see the MAI rule
14
Options
 Thesaurus
Master can be queried
to show the term record
 Broader
term
 Narrower terms
 Use For terms (“synonyms”)
 Related terms
 Scope notes
15
16
Options, continued
 MAI
can be queried to return the
rule that includes the search
word(s)
17
Show the rule
18
Options, continued
 Your
database/index of items is
then queried to bring back the
records in your collection that are
indexed with the preferred term
 For
our demo, we wrote an xquery
request into the gettitles.php file
 Our 1100-title demo records are
maintained by a MarkLogic server
19
A list of items
20
Choose the item
 Your
user clicks on the item(s)
appropriate to their query
 The
document details (or the item
itself) is returned
21
The right stuff
22
How’s it working?
What words and phrases do your
users search for?



a search log can record “misses”
a user focus group can suggest
additions
subject matter experts can help in
their area of expertise
23
Fine tuning
Modify your taxonomy to respond
to more words



add common misspellings to
rules
add alternate words as Use For
terms (synonyms) in the
thesaurus
(or as additions to the rules)
consider terms for addition to
the thesaurus (candidates)
24
The advantages
 MAIQuery
connects your user
with the controlled vocabulary
 Your user can review term
records and rulebase rules to
learn more about your taxonomy
 Your user becomes more
productive
25
Another way to connect users
 Category
search used more than
half the time for research
 Also known as directory search,
your user “drills down” from
general to specific
26
Value of Category search
 Searchers
find info 50% faster
using browsable categories than
using list returned from free text
search
 Results
even stronger when results
not in top 20 returns
 Searchers
prefer browsable
category search
Chen, H., and Dumais, S.
27
Search – the Directory Approach
28
Category: Business and Economy
29
Results: Business Libraries
30
Your Thesaurus as Directory
 Present
your controlled
vocabulary as a guide to your
collection
31
32
Thesauri OnLine

Australian Governments' Interactive
Functions Thesaurus – AGIFT
http://www.naa.gov.au/recordkeeping/thes
aurus/000482.htm

Transportation Research Thesaurus – TRT
http://ntl.bts.gov/trt/trt_topterms.jsp

NBII (National Biological Information
Infrastructure)
http://thesaurus.nbii.gov/SearchNBIIThesa
urus/about.faces
33
Presenting: TM Navtree
 Your
thesaurus presented as a
navigation aid
 User “drill down” with all the
neighboring terms visible
 Each term indicates the number
of documents indexed with it
 Terms are hyperlinks to a list of
items
34
A hierarchical tree
35
See full topic coverage by revealing
Narrower Terms
36
Choose a term
 Click
on a term, get the titles
indexed with it
37
Choose a title
 Click
on a title, get its details (or
bring up the item)
38
How it’s done
used PHP Levels, an open
source application from
SourceForge to create the tree
 An exported XML version of the
thesaurus is parsed to produce
the required text file to populate
the tree
 The content manager is queried
for the document totals
 We
39
How it’s done, continued
 When
a term is selected, it is
passed to a gettitles.php
 A bit of php code connects to the
content manager and returns a
string of data about each title
 The web page displays the data
in the format desired
40
The advantages
 TM
Navtree Top Terms describe
the organization of your
collection(s)
 Narrower terms help your user
hone in on the most appropriate
term
 Adjacent terms impart
connotation
41
The advantages
 ALL
the records indexed with the
chosen term are returned
 Your user finds what’s needed
more quickly and is more
productive
42
Questions?
Comments?
Try out the demo at
www.mediasleuth.com
See more details:
Data Harmony Programmer
Interface for Web Applications
Thank you.
Mary Garcia
43
Making Users
More Productive
MAI Query and NavTree from
Data Harmony
44