<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Archiving and Interchange DTD v1.0 20120330//EN" "JATS-archivearticle1.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink">
  <front>
    <journal-meta />
    <article-meta>
      <title-group>
        <article-title>Appendix C Results of the Domain Specific Track</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <string-name>Giorgio Maria Di Nunzio</string-name>
          <email>dinunzio@dei.unipd.it</email>
          <xref ref-type="aff" rid="aff0">0</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Nicola Ferro</string-name>
          <email>ferro@dei.unipd.it</email>
          <xref ref-type="aff" rid="aff0">0</xref>
        </contrib>
        <aff id="aff0">
          <label>0</label>
          <institution>Department of Information Engineering University of Padua Italy</institution>
        </aff>
      </contrib-group>
      <fpage>370</fpage>
      <lpage>411</lpage>
    </article-meta>
  </front>
  <body>
    <sec id="sec-1">
      <title>-</title>
      <p>Introduction
3</p>
      <p>Results for CLEF 2007 Domain Specific Track
3. Individual Experiment Results and Graphs
This section provides the individual results for each official experiment. For each experiment the following tables and graphs are
shown:
- Overall statistics and information
- Interpolated recall vs precision averages plot
- Average precision statistics and box plot
- Average precision comparison to median plot
- Document cutoff levels vs precision at DCL plot
- R-Precision statistics and box plot
- R-Precision comparison to median plot
Topics are identified with DOIs, as well as the experiments. The prefix for the DOI of a topic is 10.2452. The following example
shows how to build the DOI for a topic given its number: for topic 200-AH, the corresponding DOI is 10.2452/200-AH
List of Submitted Experiments
7</p>
    </sec>
    <sec id="sec-2">
      <title>United States 10.2415/DS-MONO-DE</title>
    </sec>
    <sec id="sec-3">
      <title>CLEF2007.CHESHIRE.BERKMDECCP15</title>
    </sec>
    <sec id="sec-4">
      <title>United States 10.2415/DS-MONO-DE</title>
    </sec>
    <sec id="sec-5">
      <title>CLEF2007.CHESHIRE.BERKMDETHP7 10.2415/DS-MONO-DECLEF2007.CHEMNITZ.CUT_DS_MONO_DE_MERGED 10.2415/DS-MONO-DE</title>
      <p>CLEF2007.CHEMNITZ.CUT_DS_MONO_DE_STRUCT
10.2415/DS-MONO-DECLEF2007.CHEMNITZ.CUT_DS_MONO_DE_UNSTRUCT
10.2415/DS-MONO-DE</p>
    </sec>
    <sec id="sec-6">
      <title>CLEF2007.UNINE.UNINEDE1</title>
      <p>10.2415/DS-MONO-DE</p>
    </sec>
    <sec id="sec-7">
      <title>CLEF2007.UNINE.UNINEDE2</title>
      <p>10.2415/DS-MONO-DE</p>
    </sec>
    <sec id="sec-8">
      <title>CLEF2007.UNINE.UNINEDE3</title>
      <p>10.2415/DS-MONO-DE</p>
    </sec>
    <sec id="sec-9">
      <title>CLEF2007.UNINE.UNINEDE4</title>
      <p>10.2415/DS-MONO-DE</p>
    </sec>
    <sec id="sec-10">
      <title>CLEF2007.XEROX.XRCELEDE</title>
      <p>10.2415/DS-MONO-DE</p>
    </sec>
    <sec id="sec-11">
      <title>CLEF2007.XEROX.XRCEPRFDE</title>
      <p>10.2415/DS-MONO-DE</p>
    </sec>
    <sec id="sec-12">
      <title>CLEF2007.XEROX.XRCEPRFDELE</title>
      <p>10.2415/DS-MONO-DE</p>
    </sec>
    <sec id="sec-13">
      <title>CLEF2007.XEROX.XRCEPRFLEDE</title>
      <p>10.2415/DS-MONO-ENCLEF2007.CHEMNITZ.CUT_DS_MONO_EN_MERGED
10.2415/DS-MONO-ENCLEF2007.CHEMNITZ.CUT_DS_MONO_EN_MERGED_S</p>
    </sec>
    <sec id="sec-14">
      <title>TRUCT</title>
      <p>10.2415/DS-MONO-ENCLEF2007.CHEMNITZ.CUT_DS_MONO_EN_MERGED_U</p>
    </sec>
    <sec id="sec-15">
      <title>NSTRUCT</title>
      <p>10.2415/DS-MONO-EN</p>
    </sec>
    <sec id="sec-16">
      <title>CLEF2007.MOSCOW.CIRDSMONOEN</title>
      <p>10.2415/DS-MONO-EN</p>
    </sec>
    <sec id="sec-17">
      <title>CLEF2007.MOSCOW.CIRDSMONOEN2 10.2415/DS-MONO-EN</title>
    </sec>
    <sec id="sec-18">
      <title>CLEF2007.UNINE.UNINEEN1</title>
      <p>10.2415/DS-MONO-EN</p>
    </sec>
    <sec id="sec-19">
      <title>CLEF2007.UNINE.UNINEEN2</title>
      <p>10.2415/DS-MONO-EN</p>
    </sec>
    <sec id="sec-20">
      <title>CLEF2007.UNINE.UNINEEN3</title>
      <p>10.2415/DS-MONO-EN</p>
    </sec>
    <sec id="sec-21">
      <title>CLEF2007.UNINE.UNINEEN4</title>
      <p>10.2415/DS-MONO-EN</p>
    </sec>
    <sec id="sec-22">
      <title>CLEF2007.XEROX.XRCELEEN</title>
      <p>10.2415/DS-MONO-EN</p>
    </sec>
    <sec id="sec-23">
      <title>CLEF2007.XEROX.XRCEPRFEN</title>
      <p>10.2415/DS-MONO-EN</p>
    </sec>
    <sec id="sec-24">
      <title>CLEF2007.XEROX.XRCEPRFENLE</title>
      <p>10.2415/DS-MONO-EN</p>
    </sec>
    <sec id="sec-25">
      <title>CLEF2007.XEROX.XRCEPRFLEEN</title>
      <p>10.2415/DS-MONO-RUCLEF2007.CHEMNITZ.CUT_DS_MONO_RU_MERGED
10.2415/DS-MONO-RUCLEF2007.CHEMNITZ.CUT_DS_MONO_RU_STRUCT
10.2415/DS-MONO-RUCLEF2007.CHEMNITZ.CUT_DS_MONO_RU_UNSTRUCT</p>
    </sec>
    <sec id="sec-26">
      <title>United States 10.2415/DS-MONO-RU</title>
    </sec>
    <sec id="sec-27">
      <title>CLEF2007.CHESHIRE.BERKMRUCCP15</title>
    </sec>
    <sec id="sec-28">
      <title>United States 10.2415/DS-MONO-RU</title>
    </sec>
    <sec id="sec-29">
      <title>CLEF2007.CHESHIRE.BERKMRUTHP7 10.2415/DS-MONO-RU</title>
    </sec>
    <sec id="sec-30">
      <title>CLEF2007.MOSCOW.CIRDSMONORU</title>
      <p>10.2415/DS-MONO-RU</p>
    </sec>
    <sec id="sec-31">
      <title>CLEF2007.MOSCOW.CIRDSMONORU2 10.2415/DS-MONO-RU</title>
    </sec>
    <sec id="sec-32">
      <title>CLEF2007.UNINE.UNINERU1</title>
      <p>10.2415/DS-MONO-RU</p>
    </sec>
    <sec id="sec-33">
      <title>CLEF2007.UNINE.UNINERU2</title>
      <p>10.2415/DS-MONO-RU</p>
    </sec>
    <sec id="sec-34">
      <title>CLEF2007.UNINE.UNINERU3</title>
      <p>10.2415/DS-MONO-RU</p>
    </sec>
    <sec id="sec-35">
      <title>CLEF2007.UNINE.UNINERU4</title>
    </sec>
    <sec id="sec-36">
      <title>DS-MONO-DE-CLEF2007</title>
      <p>chemnitz
10.2415/DS-BILI-X2DECLEF2007.CHEMNITZ.CUT_DS_BILI_EN2DE_MERGE</p>
    </sec>
    <sec id="sec-37">
      <title>D_THES</title>
      <p>10.2415/DS-BILI-X2DECLEF2007.CHEMNITZ.CUT_DS_BILI_RU2DE_MERGE
D
en
ru
ru
en
en
ru
ru
en
en
en
en
en
en
de
de
ru
de
de
ru
ru
ru
ru
de
de
de
de
de
en
de
de
en
de
de
en
en
en
en
de
en
ru</p>
      <p>TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
TD
Track Overview Results and Graphs
12
30%
20%
10%
0%</p>
      <p>0%
0.8
0.6
0.4
0.2
e
c
n
re 0
e
iff
D
−0.2
−0.4
−0.6
−0.8
−1</p>
      <sec id="sec-37-1">
        <title>Topic Identifier</title>
        <p>cheshire [Experiment BERKMDETHP7; MAP 31.99%; Pooled]
cheshire [Experiment BERKMDECCP15; MAP 31.50%; Pooled]
chemnitz [Experiment CUT_DS_MONO_DE_MERGED; MAP 29.91%; Pooled]
chemnitz [Experiment CUT_DS_MONO_DE_UNSTRUCT; MAP 28.87%; Pooled]
chemnitz [Experiment CUT_DS_MONO_DE_STRUCT; MAP 26.31%; Pooled]
0%
10%
20%
30%
70%
80%
10.2455/TUKEY_T_TEST.67D6904A15764192AAB8A24F7DD2C621</p>
        <p>Domain−Specific Monolingual German Task − Tukey T test with "top group" highlighted</p>
        <p>Domain−Specific Monolingual German Task Top 5 Participants − Retrieved documents vs Mean Precision
100%
xerox [Experiment XRCEPRFLEDE; R−Prec 50.08%; Pooled]
unine [Experiment UNINEDE4; R−Prec 39.61%; Pooled]
cheshire [Experiment BERKMDETHP7; R−Prec 35.75%; Pooled]
chemnitz [Experiment CUT_DS_MONO_DE_MERGED; R−Prec 33.84%; Pooled]
90%</p>
        <p>30 100 200
Retrieved Documents (logarithmic scale)
500
1000
1 Domain−Specific Monolingual German Task Top 5 Participants − Comparison to Median R−Precision by Topic (Topics 176−DS to 200−DS)
xerox [Experiment XRCEPRFLEDE; R−Prec 50.08%; Pooled]
unine [Experiment UNINEDE4; R−Prec 39.61%; Pooled]
cheshire [Experiment BERKMDETHP7; R−Prec 35.75%; Pooled]
chemnitz [Experiment CUT_DS_MONO_DE_MERGED; R−Prec 33.84%; Pooled]
chemnitz [Experiment CUT_DS_MONO_DE_UNSTRUCT; R−Prec 32.97%; Pooled]
chemnitz [Experiment CUT_DS_MONO_DE_STRUCT; R−Prec 30.37%; Pooled]
0%
10%
20%
30%
40%</p>
        <p>50%
R−Precision
60%
70%
80%</p>
        <p>Domain−Specific Monolingual German Task − Tukey T test with "top group" highlighted
XRCEPRFLEDE</p>
        <p>Domain−Specific Bilingual Russian Task − Distribution of the Topics of the Experiment
00% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%</p>
        <p>Average Precision
1
0.8
0.6
0.4
0.2
0%5</p>
        <p>Domain−Specific Bilingual Russian Task − Distribution of the Topics of the Experiment</p>
        <p>Domain−Specific Bilingual Russian Task − Box plot of the Topics of the Experiment</p>
        <p>Domain−Specific Bilingual Russian Task − Distribution of the Topics of the Experiment
00% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%</p>
        <p>Average Precision
1
0.8
0.6
0.4
0.2
0%5
Domain−Specific Bilingual Russian Task − Box plot of the Topics of the Experiment</p>
        <p>Domain−Specific Bilingual Russian Task − Distribution of the Topics of the Experiment
15
t
n
e
m
ir
e
p
xE10
e
h
ft
so
cp
i
o
fT 5
o
r
e
b
m
u
N</p>
        <p>Domain−Specific Multilingual Task − Box plot of the Topics of the Experiment</p>
        <p>Domain−Specific Multilingual Task − Distribution of the Topics of the Experiment
00% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%</p>
        <p>Average Precision
1
0.8
0.6
0.4
0.2
15
t
n
e
m
ir
e
p
xE10
e
h
ft
so
cp
i
o
fT 5
o
r
e
b
m
u</p>
        <p>N
−1176−DS177−DS178−DS179−DS180−DS181−DS182−DS183−DS184−DS185−DS186−DS187−DS188−DS189−DS190−DS191−DS192−DS193−DS194−DS195−DS196−DS197−DS198−DS199−DS200−DS</p>
        <sec id="sec-37-1-1">
          <title>TopicIdentifier</title>
          <p>Multilingual merged unstructured indizes translated
with Google translate (Beta), PROMT and the
provided thesaurus
0%5
Domain−Specific Multilingual Task − Box plot of the Topics of the Experiment</p>
          <p>Domain−Specific Multilingual Task − Distribution of the Topics of the Experiment
6
t
iren5
m
e
xp
E4
e
h
ftcspo3
i
o
f2
T
o
r
e
bu1
m
N
00% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%</p>
        </sec>
      </sec>
      <sec id="sec-37-2">
        <title>R−Precision</title>
        <p>−1176−DS177−DS178−DS179−DS180−DS181−DS182−DS183−DS184−DS185−DS186−DS187−DS188−DS189−DS190−DS191−DS192−DS193−DS194−DS195−DS196−DS197−DS198−DS199−DS200−DS</p>
        <sec id="sec-37-2-1">
          <title>TopicIdentifier</title>
          <p>Priority</p>
          <p>Query Construction
25,000 Source Language
9,690 Topic Fields
2,662 Pooled</p>
          <p>Domain−Specific Multilingual Task − Box plot of the Topics of the Experiment</p>
          <p>Domain−Specific Multilingual Task − Distribution of the Topics of the Experiment
00% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%
Average Precision
1
0.8
0.6
0.4
0.2
−1176−DS177−DS178−DS179−DS180−DS181−DS182−DS183−DS184−DS185−DS186−DS187−DS188−DS189−DS190−DS191−DS192−DS193−DS194−DS195−DS196−DS197−DS198−DS199−DS200−DS</p>
        </sec>
        <sec id="sec-37-2-2">
          <title>TopicIdentifier</title>
          <p>Multilingual merged unstructured indizes translated
with Google translate (Beta) and the provided
thesaurus
0%5
Domain−Specific Multilingual Task − Box plot of the Topics of the Experiment
R_PRECISION</p>
          <p>Domain−Specific Multilingual Task − Distribution of the Topics of the Experiment
6
t
iren5
m
e
xp
E4
e
h
ftscop3
i
o
f2
T
o
r
e
ub1
m
N
00% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%</p>
        </sec>
      </sec>
      <sec id="sec-37-3">
        <title>R−Precision</title>
        <p>−1176−DS177−DS178−DS179−DS180−DS181−DS182−DS183−DS184−DS185−DS186−DS187−DS188−DS189−DS190−DS191−DS192−DS193−DS194−DS195−DS196−DS197−DS198−DS199−DS200−DS</p>
        <sec id="sec-37-3-1">
          <title>TopicIdentifier</title>
          <p>Priority</p>
          <p>Query Construction
25,000 Source Language
9,690 Topic Fields
1,725 Pooled</p>
          <p>Domain−Specific Multilingual Task − Box plot of the Topics of the Experiment</p>
          <p>Domain−Specific Multilingual Task − Distribution of the Topics of the Experiment
00% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%
Average Precision
1
0.8
0.6
0.4
0.2
−0.2
−0.4
−0.6
−0.8
−1176−DS177−DS178−DS179−DS180−DS181−DS182−DS183−DS184−DS185−DS186−DS187−DS188−DS189−DS190−DS191−DS192−DS193−DS194−DS195−DS196−DS197−DS198−DS199−DS200−DS</p>
        </sec>
        <sec id="sec-37-3-2">
          <title>TopicIdentifier</title>
          <p>Domain−Specific Multilingual Task − Retrieved documents vs Mean Precision</p>
          <p>CUT_DS_MULTI_RU2X_MERGED
100%
90%
80%
70%
60%
40%
30%
20%
10%
0%5
Multilingual merged unstructured indizes translated
with Google translate (Beta), PROMT and the
provided thesaurus</p>
          <p>Domain−Specific Multilingual Task − Distribution of the Topics of the Experiment
00% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%</p>
        </sec>
      </sec>
      <sec id="sec-37-4">
        <title>R−Precision</title>
        <p>Precision averages (%) for individual
queries
−1176−DS177−DS178−DS179−DS180−DS181−DS182−DS183−DS184−DS185−DS186−DS187−DS188−DS189−DS190−DS191−DS192−DS193−DS194−DS195−DS196−DS197−DS198−DS199−DS200−DS</p>
        <sec id="sec-37-4-1">
          <title>TopicIdentifier</title>
          <p>Priority</p>
          <p>Query Construction
25,000 Source Language
9,690 Topic Fields
1,892 Pooled
100%
90%
80%
70%
60%
40%
30%
20%
10%
0%0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%</p>
        </sec>
      </sec>
      <sec id="sec-37-5">
        <title>Recall</title>
        <p>Domain−Specific Multilingual Task − Box plot of the Topics of the Experiment</p>
        <p>Domain−Specific Multilingual Task − Distribution of the Topics of the Experiment
00% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%
Average Precision
1
0.8
0.6
0.4
0.2
−0.2
−0.4
−0.6
−0.8
Precision averages (%) for individual
queries
Domain−SpecificMultilingualTask−ComparisontoMedianR−PrecisionbyTopic(Topics176−DSto200−DS)</p>
        <sec id="sec-37-5-1">
          <title>BERKMUDEP15</title>
          <p>Domain−Specific Multilingual Task − Box plot of the Topics of the Experiment
0%5</p>
          <p>Domain−Specific Multilingual Task − Distribution of the Topics of the Experiment
10
t
n
e
irm 8
e
xp
E
the 6
f
so
c
iop 4
T
f
o
rbe 2
m
u
N
00% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%</p>
        </sec>
      </sec>
      <sec id="sec-37-6">
        <title>R−Precision</title>
        <p>Overall statistics for 25 queries :
Total number of documents over all queries
Retrieved
Relevant
Relevant retrieved
Priority</p>
        <p>Query Construction
25,000 Source Language
9,690 Topic Fields
1,966 Pooled
4
AUTOMATIC
German
title, description
false
100%
90%
80%
70%
60%
40%
30%
20%
10%
0%0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%</p>
      </sec>
      <sec id="sec-37-7">
        <title>Recall</title>
        <p>Domain−Specific Multilingual Task − Box plot of the Topics of the Experiment
0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%</p>
      </sec>
      <sec id="sec-37-8">
        <title>Average Precision</title>
        <p>Domain−Specific Multilingual Task − Distribution of the Topics of the Experiment
00% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%
Average Precision
1
0.8
0.6
0.4
0.2
−0.2
−0.4
−0.6
−0.8
−1 176−DS177−DS178−DS179−DS180−DS181−DS182−DS183−DS184−DS185−DS186−DS187−DS188−DS189−DS190−DS191−DS192−DS193−DS194−DS195−DS196−DS197−DS198−DS199−DS200−DS</p>
        <sec id="sec-37-8-1">
          <title>TopicIdentifier</title>
          <p>Precision averages (%) for individual
queries
Domain−SpecificMultilingualTask−ComparisontoMedianR−PrecisionbyTopic(Topics176−DSto200−DS)</p>
        </sec>
        <sec id="sec-37-8-2">
          <title>BERKMUDETHP7</title>
          <p>Domain−Specific Multilingual Task − Box plot of the Topics of the Experiment
20%
10%
0%5
−0.2
−0.4
−0.6
−0.8</p>
          <p>Domain−Specific Multilingual Task − Distribution of the Topics of the Experiment
00% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%</p>
        </sec>
      </sec>
      <sec id="sec-37-9">
        <title>R−Precision</title>
        <p>−1 176−DS177−DS178−DS179−DS180−DS181−DS182−DS183−DS184−DS185−DS186−DS187−DS188−DS189−DS190−DS191−DS192−DS193−DS194−DS195−DS196−DS197−DS198−DS199−DS200−DS</p>
        <sec id="sec-37-9-1">
          <title>TopicIdentifier</title>
          <p>Overall statistics for 25 queries :
Total number of documents over all queries
Retrieved
Relevant
Relevant retrieved
Priority</p>
          <p>Query Construction
25,000 Source Language
9,690 Topic Fields
3,323 Pooled
100%
90%
80%
70%
60%
40%
30%
20%
10%
0%0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%</p>
        </sec>
      </sec>
      <sec id="sec-37-10">
        <title>Recall</title>
        <p>Domain−Specific Multilingual Task − Box plot of the Topics of the Experiment
0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%</p>
      </sec>
      <sec id="sec-37-11">
        <title>Average Precision</title>
        <p>Domain−Specific Multilingual Task − Distribution of the Topics of the Experiment
00% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%
Average Precision
1
0.8
0.6
0.4
0.2
−0.2
−0.4
−0.6
−0.8
Docs Cutoff Levels Precision at DCL (%)
5 docs 24.80
10 docs 23.60
15 docs 26.93
20 docs 28.60
30 docs 27.87
100 docs 27.80
200 docs 24.58
500 docs 18.44
1000 docs 13.29
R-Precision (precision after R document retrieved,
where R = Relevant retrieved)
18.52</p>
        <p>Domain−Specific Multilingual Task − Box plot of the Topics of the Experiment
0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%</p>
      </sec>
      <sec id="sec-37-12">
        <title>R−Precision</title>
        <p>Domain−Specific Multilingual Task − Distribution of the Topics of the Experiment
5
t
n
e
ie4
m
r
xp
E
fth3
e
so
c
ip2
o
T
f
o
r
e
b1
m
u
N
00% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%</p>
      </sec>
      <sec id="sec-37-13">
        <title>R−Precision</title>
        <p>−0.2
−0.4
−0.6
−0.8
Precision averages (%) for individual
queries
100%
90%
80%
70%
60%
40%
30%
20%
10%
0%5
Overall statistics for 25 queries :
Total number of documents over all queries
Retrieved
Relevant
Relevant retrieved
Priority</p>
        <p>Query Construction
25,000 Source Language
9,690 Topic Fields
3,237 Pooled
5
AUTOMATIC
English
title, description
false
100%
90%
80%
70%
60%
40%
30%
20%
10%
0%0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%</p>
      </sec>
      <sec id="sec-37-14">
        <title>Recall</title>
        <p>Domain−Specific Multilingual Task − Box plot of the Topics of the Experiment
0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%</p>
      </sec>
      <sec id="sec-37-15">
        <title>Average Precision</title>
        <p>Domain−Specific Multilingual Task − Distribution of the Topics of the Experiment
00% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%
Average Precision
1
0.8
0.6
0.4
0.2
ce
reen 0
if
D
−0.2
−0.4
−0.6
−0.8
−1 176−DS177−DS178−DS179−DS180−DS181−DS182−DS183−DS184−DS185−DS186−DS187−DS188−DS189−DS190−DS191−DS192−DS193−DS194−DS195−DS196−DS197−DS198−DS199−DS200−DS</p>
        <sec id="sec-37-15-1">
          <title>TopicIdentifier</title>
          <p>Docs Cutoff Levels Precision at DCL (%)
5 docs 24.00
10 docs 23.60
15 docs 28.27
20 docs 28.20
30 docs 27.47
100 docs 26.64
200 docs 23.62
500 docs 17.80
1000 docs 12.95
R-Precision (precision after R document retrieved,
where R = Relevant retrieved)
17.84</p>
          <p>Domain−Specific Multilingual Task − Retrieved documents vs Mean Precision</p>
          <p>Domain−Specific Multilingual Task − Box plot of the Topics of the Experiment
0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%</p>
        </sec>
      </sec>
      <sec id="sec-37-16">
        <title>R−Precision</title>
        <p>Domain−Specific Multilingual Task − Distribution of the Topics of the Experiment
5
t
n
e
ie4
m
r
xp
E
tfh3
e
so
c
ip2
o
T
f
o
r
e
b1
m
u
N
00% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%</p>
      </sec>
      <sec id="sec-37-17">
        <title>R−Precision</title>
        <p>2.59
19.49
11.00
7.25
39.11
18.45
36.34
0.00
4.46
34.63
8.86
17.73
1
0.8
0.6
0.4
0.2
ce
reen 0
if
D
−0.2
−0.4
−0.6
−0.8
Precision averages (%) for individual
queries
−1 176−DS177−DS178−DS179−DS180−DS181−DS182−DS183−DS184−DS185−DS186−DS187−DS188−DS189−DS190−DS191−DS192−DS193−DS194−DS195−DS196−DS197−DS198−DS199−DS200−DS</p>
        <p>TopicIdentifier
100%
90%
80%
70%
60%
40%
30%
20%
10%
0%5
Overall statistics for 25 queries :
Total number of documents over all queries
Retrieved
Relevant
Relevant retrieved
Priority</p>
        <p>Query Construction
25,000 Source Language
9,690 Topic Fields
2,125 Pooled
0.0167 Multilingual from Russian using Powertrans and
0.1259 Class Cluster Lookup
100%
90%
80%
70%
60%
40%
30%
20%
10%
0%0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%</p>
      </sec>
      <sec id="sec-37-18">
        <title>Recall</title>
        <p>Domain−Specific Multilingual Task − Box plot of the Topics of the Experiment
0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%</p>
      </sec>
      <sec id="sec-37-19">
        <title>Average Precision</title>
        <p>Domain−Specific Multilingual Task − Distribution of the Topics of the Experiment
00% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%
Average Precision
1
0.8
0.6
0.4
0.2
ce
reen 0
if
D
−0.2
−0.4
−0.6
−0.8
20
t
n
e
m
i
rpe15
x
E
e
h
ftso10
cp
i
o
T
f
o
re 5
b
m
u</p>
        <p>N
−1 176−DS177−DS 178−DS179−DS180−DS181−DS182−DS 183−DS184−DS185−DS186−DS187−DS 188−DS189−DS190−DS191−DS192−DS 193−DS194−DS195−DS196−DS197−DS 198−DS199−DS200−DS</p>
        <sec id="sec-37-19-1">
          <title>TopicIdentifier</title>
          <p>Docs Cutoff Levels Precision at DCL (%)
5 docs 22.40
10 docs 17.60
15 docs 17.33
20 docs 16.20
30 docs 15.60
100 docs 15.44
200 docs 14.44
500 docs 11.29
1000 docs 8.50
R-Precision (precision after R document retrieved,
where R = Relevant retrieved)
12.38</p>
          <p>Domain−Specific Multilingual Task − Retrieved documents vs Mean Precision</p>
          <p>Domain−Specific Multilingual Task − Box plot of the Topics of the Experiment
0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%</p>
        </sec>
      </sec>
      <sec id="sec-37-20">
        <title>R−Precision</title>
        <p>Domain−Specific Multilingual Task − Distribution of the Topics of the Experiment
8
tn7
e
m
ire6
xp
E
ftscope45
h
i
o3
T
f
o
r2
e
b
m
u1
N
00% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%</p>
      </sec>
      <sec id="sec-37-21">
        <title>R−Precision</title>
        <p>0.00
27.94
18.00
8.70
13.87
16.14
17.40
4.23
11.14
12.16
0.63
13.04
1
0.8
0.6
0.4
0.2
ce
reen 0
if
D
−0.2
−0.4
−0.6
−0.8
Precision averages (%) for individual
queries
−1 176−DS 177−DS 178−DS 179−DS 180−DS 181−DS 182−DS 183−DS 184−DS 185−DS 186−DS 187−DS 188−DS 189−DS 190−DS 191−DS 192−DS 193−DS 194−DS 195−DS 196−DS 197−DS 198−DS 199−DS 200−DS</p>
        <p>TopicIdentifier
100%
90%
80%
70%
60%
40%
30%
20%
10%
0%5
Overall statistics for 25 queries :
Total number of documents over all queries
Retrieved
Relevant
Relevant retrieved
Priority</p>
        <p>Query Construction
25,000 Source Language
9,690 Topic Fields
2,087 Pooled
6
AUTOMATIC
Russian
title, description
false
0.0156 Multilingual from Russian using Powertrans and
0.1230 Thesaurus Lookup
100%
90%
80%
70%
60%
40%
30%
20%
10%
0%0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%</p>
      </sec>
      <sec id="sec-37-22">
        <title>Recall</title>
        <p>Domain−Specific Multilingual Task − Box plot of the Topics of the Experiment
0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%</p>
      </sec>
      <sec id="sec-37-23">
        <title>Average Precision</title>
        <p>Domain−Specific Multilingual Task − Distribution of the Topics of the Experiment
00% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%
Average Precision
1
0.8
0.6
0.4
0.2
ce
reen 0
if
D
−0.2
−0.4
−0.6
−0.8
20
t
n
e
m
i
rpe15
x
E
e
h
tfso10
cp
i
o
T
f
o
re 5
b
m
u</p>
        <p>N
−1 176−DS177−DS178−DS179−DS180−DS181−DS182−DS183−DS184−DS185−DS186−DS187−DS188−DS189−DS190−DS191−DS192−DS193−DS194−DS195−DS196−DS197−DS198−DS199−DS200−DS</p>
        <sec id="sec-37-23-1">
          <title>TopicIdentifier</title>
          <p>0.04
9.89
5.53
1.34
2.70
5.57
4.19
0.37
1.70
3.31
0.03
2.81
Precision averages (%) for individual
queries
Domain−Specific Multilingual Task − Retrieved documents vs Mean Precision
Domain−Specific Multilingual Task − Box plot of the Topics of the Experiment
100%
90%
80%
70%
60%
40%
30%
20%
10%
Interquartile range
Mean
Mean With No Outliers
Std With No Outliers
176-DS
177-DS
178-DS
179-DS
180-DS
181-DS
182-DS
183-DS
184-DS
185-DS
186-DS
187-DS
188-DS
0.55
2.26
1.64
29.31
24.34
22.58
2.96
10.25
3.44
9.48
3.03
22.97
27.92
189-DS
190-DS
191-DS
192-DS
193-DS
194-DS
195-DS
196-DS
197-DS
198-DS
199-DS
200-DS
1</p>
        </sec>
      </sec>
    </sec>
  </body>
  <back>
    <ref-list>
      <ref id="ref1">
        <mixed-citation>
          <source>0.0073 Multilingual from German using Powertrans and Class 0.1166 Cluster Lookup 0.0085 Multilingual from German using Powertrans and 0.1214 Thesaurus Lookup 0.0484 Multilingual from English using Powertrans and 0</source>
          .
          <source>1947 Class Cluster Lookup 0.0435 Multilingual from English using Powertrans and 0</source>
          .1916 Thesaurus Lookup
        </mixed-citation>
      </ref>
    </ref-list>
  </back>
</article>