=Paper=
{{Paper
|id=Vol-1173/CLEF2007wn-GeoCLEF-Larson2007
|storemode=property
|title=Cheshire at GeoCLEF 2007: Retesting Text Retrieval Baselines
|pdfUrl=https://ceur-ws.org/Vol-1173/CLEF2007wn-GeoCLEF-Larson2007.pdf
|volume=Vol-1173
|dblpUrl=https://dblp.org/rec/conf/clef/Larson07b
}}
==Cheshire at GeoCLEF 2007: Retesting Text Retrieval Baselines==
https://ceur-ws.org/Vol-1173/CLEF2007wn-GeoCLEF-Larson2007.pdf
Cheshire at GeoCLEF 2007: Retesting Text
Retrieval Baselines
Ray R. Larson
School of Information
University of California, Berkeley, USA
ray@sims.berkeley.edu
Abstract
In this paper we will briefly describe the approaches taken by Berkeley for the main
GeoCLEF 2007 tasks (Mono and Bilingual retrieval). This year we used only a single
system in the research, and were not able to do much in the way of interesting geo-
graphic work due to a number of factors, not the least of which was time competition
from other tasks. The approach this year used probabilistic text retrieval based on
logistic regression and incorporating blind relevance feedback for all of the runs. All
translation for bilingual tasks was performed using the LEC Power Translator PC-
based MT system. Since data on overall performance relative to other participants
were not available at the time of writing, our discussion in this paper is limited to
comparison between our submitted runs.
Categories and Subject Descriptors
H.3 [Information Storage and Retrieval]: H.3.1 Content Analysis and Indexing; H.3.3 Infor-
mation Search and Retrieval
General Terms
Algorithms, Performance, Measurement
Keywords
Cheshire II, Logistic Regression
1 Introduction
This paper describes the retrieval algorithms and evaluation results for Berkeley’s official sub-
missions for the GeoCLEF 2007 track. Instead of the expansion approaches used in last year’s
GeoCLEF, this year we used only unexpanded topics in querying the database. All of the runs
were automatic without manual intervention in the queries (or translations). We submitted 3
Monolingual runs (1 German, 1 English, and 1 Portuguese) and 9 Bilingual runs (each of the
three main languages to each each other language, and 3 runs from Spanish to English, German,
and Portuguese).
This paper first describes the retrieval algorithms used for our submissions, followed by a
discussion of the processing used for the runs. We then examine the results obtained for our
official runs, and finally present conclusions and future directions for GeoCLEF participation.
2 The Retrieval Algorithms
Note that this section is virtually identical to one that appears in our ImageCLEF and Domain
Specific papers. The basic form and variables of the Logistic Regression (LR) algorithm used for
all of our submissions was originally developed by Cooper, et al. [5]. As originally formulated, the
LR model of probabilistic IR attempts to estimate the probability of relevance for each document
based on a set of statistics about a document collection and a set of queries in combination
with a set of weighting coefficients for those statistics. The statistics to be used and the values
of the coefficients are obtained from regression analysis of a sample of a collection (or similar
test collection) for some set of queries where relevance and non-relevance has been determined.
More formally, given a particular query and a particular document in a collection P (R | Q, D)
is calculated and the documents or components are presented to the user ranked in order of
decreasing values of that probability. To avoid invalid probability values, the usual calculation of
P (R | Q, D) uses the “log odds” of relevance given a set of S statistics, si , derived from the query
and database, such that:
XS
log O(R | Q, D) = b0 + b i si (1)
i=1
where b0 is the intercept term and the bi are the coefficients obtained from the regression analysis of
the sample collection and relevance judgements. The final ranking is determined by the conversion
of the log odds form to probabilities:
elog O(R|Q,D)
P (R | Q, D) = (2)
1 + elog O(R|Q,D)
2.1 TREC2 Logistic Regression Algorithm
For GeoCLEF we used a version the Logistic Regression (LR) algorithm that has been used very
successfully in Cross-Language IR by Berkeley researchers for a number of years[3]. The formal
definition of the TREC2 Logistic Regression algorithm used is:
p(R|C, Q) p(R|C, Q)
log O(R|C, Q) = log = log
1 − p(R|C, Q) p(R|C, Q)
|Qc |
1 X qtfi
= c0 + c1 ∗ p
|Qc | + 1 i=1 ql + 35
|Qc |
1 X tfi
+ c2 ∗ p log (3)
|Qc | + 1 i=1 cl + 80
|Qc |
1 X ctfi
− c3 ∗ p log
|Qc | + 1 i=1 Nt
+ c4 ∗ |Qc |
where C denotes a document component (i.e., an indexed part of a document which may be the
entire document) and Q a query, R is a relevance variable,
p(R|C, Q) is the probability that document component C is relevant to query Q,
p(R|C, Q) the probability that document component C is not relevant to query Q, which is 1.0 -
p(R|C, Q)
|Qc | is the number of matching terms between a document component and a query,
qtfi is the within-query frequency of the ith matching term,
tfi is the within-document frequency of the ith matching term,
ctfi is the occurrence frequency in a collection of the ith matching term,
ql is query length (i.e., number of terms in a query like |Q| for non-feedback situations),
cl is component length (i.e., number of terms in a component), and
Nt is collection length (i.e., number of terms in a test collection).
ck are the k coefficients obtained though the regression analysis.
If stopwords are removed from indexing, then ql, cl, and Nt are the query length, document
length, and collection length, respectively. If the query terms are re-weighted (in feedback, for
example), then qtfi is no longer the original term frequency, but the new weight, and ql is the
sum of the new weight values for the query terms. Note that, unlike the document and collection
lengths, query length is the “optimized” relative frequency without first taking the log over the
matching terms.
The coefficients were determined by fitting the logistic regression model specified in log O(R|C, Q)
to TREC training data using a statistical software package. The coefficients, ck , used for our of-
ficial runs are the same as those described by Chen[1]. These were: c0 = −3.51, c1 = 37.4,
c2 = 0.330, c3 = 0.1937 and c4 = 0.0929. Further details on the TREC2 version of the Logistic
Regression algorithm may be found in Cooper et al. [4].
2.2 Blind Relevance Feedback
In addition to the direct retrieval of documents using the TREC2 logistic regression algorithm
described above, we have implemented a form of “blind relevance feedback” as a supplement to the
basic algorithm. The algorithm used for blind feedback was originally developed and described by
Chen [2]. Blind relevance feedback has become established in the information retrieval community
due to its consistent improvement of initial search results as seen in TREC, CLEF and other
retrieval evaluations [6]. The blind feedback algorithm is based on the probabilistic term relevance
weighting formula developed by Robertson and Sparck Jones [8].
Blind relevance feedback is typically performed in two stages. First, an initial search using
the original topic statement is performed, after which a number of terms are selected from some
number of the top-ranked documents (which are presumed to be relevant). The selected terms
are then weighted and then merged with the initial query to formulate a new query. Finally the
reweighted and expanded query is submitted against the same collection to produce a final ranked
list of documents. Obviously there are important choices to be made regarding the number of
top-ranked documents to consider, and the number of terms to extract from those documents. For
ImageCLEF this year, having no prior data to guide us, we chose to use the top 10 terms from 10
top-ranked documents. The terms were chosen by extracting the document vectors for each of the
10 and computing the Robertson and Sparck Jones term relevance weight for each document. This
weight is based on a contingency table where the counts of 4 different conditions for combinations
of (assumed) relevance and whether or not the term is, or is not in a document. Table 1 shows
this contingency table.
Relevant Not Relevant
In doc Rt Nt − Rt Nt
Not in doc R − Rt N − N t − R + Rt N − Nt
R N −R N
Table 1: Contingency table for term relevance weighting
The relevance weight is calculated using the assumption that the first 10 documents are relevant
and all others are not. For each term in these documents the following weight is calculated:
Rt
R−Rt
wt = log Nt −Rt
(4)
N −Nt −R+Rt
The 10 terms (including those that appeared in the original query) with the highest wt are
selected and added to the original query terms. For the terms not in the original query, the new
“term frequency” (qtfi in main LR equation above) is set to 0.5. Terms that were in the original
query, but are not in the top 10 terms are left with their original qtfi . For terms in the top 10 and
in the original query the new qtfi is set to 1.5 times the original qtfi for the query. The new query
is then processed using the same LR algorithm as shown in Equation 4 and the ranked results
returned as the response for that topic.
3 Approaches for GeoCLEF
In this section we describe the specific approaches taken for our submitted runs for the GeoCLEF
task. First we describe the indexing and term extraction methods used, and then the search
features we used for the submitted runs.
3.1 Indexing and Term Extraction
The Cheshire II system uses the XML structure of the documents to extract selected portions for
indexing and retrieval. Any combination of tags can be used to define the index contents.
Table 2: Cheshire II Indexes for GeoCLEF 2006
Name Description Content Tags Used
docno Document ID DOCNO no
pauthor Author Names BYLINE, AU no
headline Article Title HEADLINE, TITLE, LEAD, LD, TI no
topic Content Words HEADLINE, TITLE, TI, LEAD yes
BYLINE, TEXT, LD, TX yes
date Date of Publication DATE, WEEK no
geotext Validated place names TEXT, LD, TX no
geopoint Validated coordinates TEXT, LD, TX no
for place names
geobox Validated bounding boxes TEXT, LD, TX no
for place names
Table 2 lists the indexes created by the Cheshire II system for the GeoCLEF database and the
document elements from which the contents of those indexes were extracted. The “Used” column
in Table 2 indicates whether or not a particular index was used in the submitted GeoCLEF runs.
The georeferencing indexing subsystem of Cheshire II was used for the geotext, geopoint, and
geobox indexes. This subsystem is intended to extract proper nouns from the text being indexed
and then attempts to match them in a digital gazetteer. For GeoCLEF we used a gazetteer
derived from the World Gazetteer (http://www.world-gazetteer.com) with 224698 entries in both
English and German. The indexing subsystem provides three different index types: verified place
names (an index of names which matched the gazetteer), point coordinates (latitude and longitude
coordinates of the verified place name) and bounding box coordinates (bounding boxes for the
matched places from the gazetteer). All three types were created, but due to time constraints,
and lack of time to upgrade the gazetteer or fix bugs in coordinate assignments, ended up not
using any of the geographic indexes in this year’s submissions. Because we do not use complete
NLP parsing techniques, the system is unable to distinguish between proper nouns for places from
Table 3: Submitted GeoCLEF Runs
Run Name Description Type MAP
BerkMODEBASE Monolingual German TD auto 0.1392
BerkMOENBASE* Monolingual English TD auto 0.2642
BerkMOPTBASE Monolingual Portuguese TD auto 0.1739
BerkBIENDEBASE Bilingual English⇒German TD auto 0.0902
BerkBIENPTBASE Bilingual English⇒Portuguese TD auto 0.2012
BerkBIDEENBASE* Bilingual German⇒English TD auto 0.2208
BerkBIDEPTBASE Bilingual German⇒Portuguese TD auto 0.0915
BerkBIPTDEBASE Bilingual Portuguese⇒German TD auto 0.1109
BerkBIPTENBASE Bilingual Portuguese⇒English TD auto 0.2112
BerkBIESDEBASE Bilingual Spanish⇒German TD auto 0.0724
BerkBIESENBASE Bilingual Spanish⇒English TD auto 0.2195
BerkBIESPTBASE Bilingual Spanish⇒Portuguese TD auto 0.1924
those for individuals. This leads to errors in geographic assignment where, for example, articles
about Irving Berlin might be tagged as refering to the city.
Because there was no explicit tagging of location-related terms in the collections used for
GeoCLEF, we applied the above approach to the “TEXT”, “LD”, and “TX” elements of the records
of the various collections. The part of news articles normally called the “dateline” indicating the
location of the news story was not separately tagged in any of the GeoCLEF collections, but often
appeared as the first part of the text for the story.
Geographic indexes were not created for the Portuguese sub-collection due to the lack of a
suitable gazetteer. We plan for later work to substitute the “GeoNames” database which is much
more detailed and provides a more complete geographical hierarchy in its records, along with
alternate names in multiple languages.
For all indexing we used language-specific stoplists to exclude function words and very common
words from the indexing and searching. The German language runs did not use decompounding
in the indexing and querying processes to generate simple word forms from compounds. Although
we tried again this year to make this work within the Cheshire system, we again lacked the time
needed to implement it correctly.
The Snowball stemmer was used by Cheshire for language-specific stemming.
3.2 Search Processing
Searching the GeoCLEF collection using the Cheshire II system involved using TCL scripts to
parse the topics and submit the title and description or the title, description, and narrative from
the topics. For monolingual search tasks we used the topics in the appropriate language (English,
German, and Portuguese), for bilingual tasks the topics were translated from the source language
to the target language using the LEC Power Translator PC-based machine translation system. In
all cases the “title” and “desc” topic elements were combined into a single probabilistic query. We
consider all of these runs to be the simplest “baseline” for our system, and we plan to implement
more elaborate processing approaches for subsequent testing.
4 Results for Submitted Runs
The summary results (as Mean Average Precision) for the submitted bilingual and monolingual
runs for both English and German are shown in Table 3, the Recall-Precision curves for these
runs are also shown in Figures 1 (for monolingual) and 2 (for bilingual). In Figures 1 and 2 the
Figure 1: Berkeley Monolingual Runs – English (left) and German (right)
1
DE
0.9 EN
0.8 PT
0.7
Precision
0.6
0.5
0.4
0.3
0.2
0.1
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Recall
TASK MAP 2006 MAP 2007 Pct. Diff.
Monolingual English 0.2499 0.2642 5.7222
Monolingual German 0.2151 0.1392 -54.5259
Monolingual Portuguese 0.1622 0.1739 7.2133
Bilingual English⇒German 0.1561 0.0902 -73.0599
Bilingual English⇒Portuguese 0.12603 0.2012 59.6825
Table 4: Comparison of Berkeley’s best 2005 and 2006 runs for English and German
names for the individual runs represent the language codes, which can easily be compared with
full names and descriptions in Table 3 (since each language combination has only a single run).
Table 3 indicates runs that had the highest overall MAP for the task by asterisks next to the
run name.
Once again we found some rather anomalous results among the official runs. For example,
it is not at all clear, given the same basic approach used for all of the runs, why the bilingual
runs for English⇒Portuguese (MAP 0.2012), and Spanish⇒Portuguese (MAP 0.1924) should have
performed better than our Monolingual Portuguese run (MAP 0.1739).
Obviously the “weak man” in our current implementation is German. This may be due to
decompounding issues, but the lower results are clear in both Monolingual and Bilingual runs
where either the source topics or the target data is German.
5 Conclusions
Although we did not do any explicit geographic processing for this year, we plan to do so in the
future. The challenge for next year is to be able to obtain the kind of effectiveness improvement
seen with manual query expansion, in automatic queries using geographic processing. In addition,
we used only the title and desc elements of topics this year, and also we did not use automatic
expansion of toponyms in the topic texts. Since this was done explicitly in some of the topic
narratives we may have missed possible improvements by not using the entire topic. In previous
years it has been apparent that implicit or explicit toponym inclusion in queries, as might be
Figure 2: Berkeley Bilingual Runs – To German (top left), To English (top right) and to Portuguese
(lower)
1 1
EN-DE DE-EN
0.9 ES-DE 0.9 ES-EN
0.8 PT-DE 0.8 PT-EN
0.7 0.7
Precision
Precision
0.6 0.6
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0 0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Recall Recall
1
EN-PT
0.9 ES-PT
0.8 DE-PT
0.7
Precision
0.6
0.5
0.4
0.3
0.2
0.1
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Recall
expected, leads to better performance when compared to using titles and descriptions alone in
retrieval.
Because we used a virtually identical processing approach (except for translation) this year
as we used for some of our runs submitted for GeoCLEF 2006, we build Table 4 examine the
differences. Overall, we did see some improvements in results. However, the submitted 2006
results used decompounding for German, which would appear to be the primary cause of our
declining monolingual and bilingual scores for German, although the translation software may
also be at fault. Otherwise, our bilingual results this year are largely due to the effectiveness
of our new translation software. We used the Spanish topic statements provided for bilingual
Spanish to English, German, and Portuguese, and saw results that look quite good for English
and Portuguese, with the exception again being German. We will be interested to see how these
scores compare to the various other approaches used in GeoCLEF this year.
References
[1] Aitao Chen. Multilingual information retrieval using english and chinese queries. In Carol
Peters, Martin Braschler, Julio Gonzalo, and Michael Kluck, editors, Evaluation of Cross-
Language Information Retrieval Systems: Second Workshop of the Cross-Language Evaluation
Forum, CLEF-2001, Darmstadt, Germany, September 2001, pages 44–58. Springer Computer
Scinece Series LNCS 2406, 2002.
[2] Aitao Chen. Cross-Language Retrieval Experiments at CLEF 2002, pages 28–48. Springer
(LNCS #2785), 2003.
[3] Aitao Chen and Fredric C. Gey. Multilingual information retrieval using machine translation,
relevance feedback and decompounding. Information Retrieval, 7:149–182, 2004.
[4] W. S. Cooper, A. Chen, and F. C. Gey. Full Text Retrieval based on Probabilistic Equations
with Coefficients fitted by Logistic Regression. In Text REtrieval Conference (TREC-2), pages
57–66, 1994.
[5] William S. Cooper, Fredric C. Gey, and Daniel P. Dabney. Probabilistic retrieval based on
staged logistic regression. In 15th Annual International ACM SIGIR Conference on Research
and Development in Information Retrieval, Copenhagen, Denmark, June 21-24, pages 198–210,
New York, 1992. ACM.
[6] Ray R. Larson. Probabilistic retrieval, component fusion and blind feedback for XML retrieval.
In INEX 2005, pages 225–239. Springer (Lecture Notes in Computer Science, LNCS 3977),
2006.
[7] Ray R. Larson, Fredric C. Gey, and Vivien Petras. Berkeley at GeoCLEF: Logistic regression
and fusion for geographic information retrieval. In Cross-Language Evaluation Forum: CLEF
2005, pages 963–976. Springer (Lecture Notes in Computer Science LNCS 4022), 2006.
[8] S. E. Robertson and K. Sparck Jones. Relevance weighting of search terms. Journal of the
American Society for Information Science, pages 129–146, May–June 1976.