ArticlePDF Available

User Behavior and Evaluation of Multilingual Information Access in Digital Libraries


Abstract and Figures

While the importance of multilingual access to information systems is undoubted, few truly operational systems exist and can serve as examples. This dissertation addresses the issue of what the user expectations and the consequences for system development are in a multilingual information environment. It starts with a general overview over the aspects of multilingual access in digital libraries. Building on previous experiences, the study focuses on a combination of log file analysis and an usability test on user needs and desired features for multilingual access based on a functional digital library with multilingual requirements (Europeana). I present the Europeana Clickstream Logger, which logs and gathers extended information on user behavior, and show first examples of the data collection possibilities. The outcome of the analysis is a description of user requirements. The dissertation concludes with the development of a possible approach for the design of multilingual information systems.
Content may be subject to copyright.
User Behavior and Evaluation of Multilingual
Information Access in Digital Libraries
Maria Gäde
Berlin School of Library and Information Science,
Dorotheenstr. 26, 10117 Berlin, Germany
Abstract. While the importance of multilingual access to information systems
is undoubted, few truly operational systems exist and can serve as examples.
This dissertation addresses the issue of what the user expectations and the
consequences for system development are in a multilingual information
environment. It starts with a general overview over the aspects of multilingual
access in digital libraries. Building on previous experiences, the study focuses
on a combination of log file analysis and an usability test on user needs and
desired features for multilingual access based on a functional digital library
with multilingual requirements (Europeana). I present the Europeana
Clickstream Logger, which logs and gathers extended information on user
behavior, and show first examples of the data collection possibilities. The
outcome of the analysis is a description of user requirements. The dissertation
concludes with the development of a possible approach for the design of
multilingual information systems.
Keywords: CLIR, MLIA, user study, Log file analysis
1 Introduction
Most of the world´s people have a native tongue other than English. In contrast, more
than 70% of the public web sites are expressed in English1. More and more users
need support to retrieve relevant information across languages boundaries [16].
Especially digital libraries, such as Europeana2, need to provide methods and tools
that enable people to access multilingual information more effectively.
Increasingly, research is concerned with requirements for and the development of
multilingual information systems. Multilingual information access (MLIA), as it is
used in the dissertation, includes all issues of accessing, searching and retrieving data
irrespective of the language in which information objects are expressed [17][18].
Cross-language information retrieval (CLIR) technologies are targeted on answering
2 Maria Gäde
queries in one language with a list of objects in other languages. This can be achieved
either by query or/and document translation, whereas the query translation usually is
The research project reported in this paper focuses on the user side and aims to
understand search processes, including users’ interaction with multilingual digital
libraries. The study addresses the issue of how users behave and interact and what the
consequences for system development in a multilingual information environment are.
The paper is structured as follows: Chapter 2 gives an overview of the different
levels and functions of MLIA systems. Chapter 3 discusses related work and the main
findings of different studies, which will be summarized and included in the
evaluation. The proposed research, focused on transaction log analysis (TLA) and
an usability test is described in Chapter 4. Following in Chapter 5, the connection
between user requirements and system requirements is shown. I conclude with a short
summary and open questions to be discussed on the consortium.
2 Aspects of Multilingual Information Access in Digital
There are different levels of multilinguality the user is confronted with. This chapter
gives an overview of varied levels and functions of search based systems such as
Europeana. The European Digital Library will provide a multilingual common access
to Europe´s cultural heritage.
Following, several aspects of MLIA regarding interface issues, the input of search
terms and the display of the results are presented.
2.1 Multilingual User Interface
The most elementary level of multilinguality is the user interface. The translation of
all static content elements on the information system’s publicly viewable web sites
and a systematic administration of language information for all content elements is
called “language-skinning”.
Currently two different options for language determination are available:
1. The user selects the interface language by a drop-down-menu or logos (e.g. flag
2. The language interface is selected automatically based on the language settings of
the user agent (i.e. browser) or the geographic location of the user determined via
User Behavior and Evaluation of Multilingual Information Access in Digital Libraries
2.2 Multilingual Search
The most essential component of a truly multilingual information system is the
multilingual search function. Interactive MLIA systems provide an additional
challenge to designers, because users may not have the necessary language skills to
find and interact with objects written in multiple languages. To provide effective
access to multilingual document collections, users require search assistance. Three
approaches for multilingual search capabilities exist today:
1. Query translation: the original query is translated into additional languages that the
document collection contains
2. Document translation: the documents in the collection are translated into the query
3. Interlingua: both queries and documents are translated into a single language,
which transforms the multilingual information retrieval process to a monolingual
The query translation process includes several stages such as query formulation
and reformulation, language detection, and translation which posit particularly
challenges for MLIA systems. The disambiguation of terms is even more problematic
when more than one language is used.
Regarding multilingual search functionalities it needs to be clarified what kind of
interaction is desired and useful to achieve optimal query translation and how systems
can help the user select the most appropriate translations, especially with ambiguous
2.3 Multilingual Result Representation and Filtering
The multilingual result representation can be performed at two levels: at the metadata
or the digital objects level. . For textual documents, it needs to be determined whether
result translation happens at the metadata level or the original document level. Within
metadata records, the most appropriate translation candidates are titles and subject
The possibility to filter a result set by language determining can usually be
implemented in two ways:
Advanced Search: a user can determine the desired language of the documents in
the result set by choosing from a list of available languages.
Refinement filter for result set: the user can filter a result set by language after the
first search has been processed.
How to present results in different languages is still an open question and needs
further research [17].
4 Maria Gäde
3 User Preferences for Multilingual Information Access in
Digital Libraries
In line with their efforts on establishing multilingual access to their content, several
digital libraries have conducted studies on user needs and requirements, but few have
paid specific attention to multilingual issues. The studies use a variety of research
methods, including observation, surveys, interviews, experiments, and transaction log
analysis. Although similar methods were used, a comparison or generalization of
these findings is difficult because of the very different user groups involved. Some
researchers selected their participants with regard to their information needs, others
with regard to their language skills. The number of participants varies considerably
between the different studies.
Since 2000 the Cross Language Evaluation Forum3 (CLEF) has carried out several
experiments with cross-language search tasks. Especially the interactive track, iCLEF,
focused on problems of multilingual search assistance and the LogCLEF track 2009
provided interesting approaches [2][9][15]. One of the most basic outcomes of the
experiments is that support for user-assisted translation of the query improves search
results [24]. Additionally the TrebleCLEF Coordination Action4 organized a
“Workshop on Best Practices for the Development of Multilingual Information
Access Systems: the User Perspective” [25] to identify the essential features that
MLIA systems should offer. The report presents some general requirements and best
practice recommendations which were collected from experiments in iCLEF [24].
iCLEF 2008 as well as studies with the Google web search engine and the Google
Translate service analyzed the behavior of users when facing strictly multilingual
information access task in order to identify the differences in the search behavior
according to the language skills [1] [20].
The European Digital Library (TEL)5 and the EDL project also prepared user
surveys and log file analysis to gather user requirements. They analyzed weblogs, the
Gabriel guestbook and the Gabriel search engine queries [12], with the result, that
multilingualism is one of the biggest problems in accessing portals. The full
translation of documents is not required, only subject translation seems to be useful.
Most users are satisfied to have the possibility to decide whether a document is
relevant or not. [25].
The University of Padua analyzed the IIS http traffic logs of The European Library
portal. One main characteristic about the sessions was that 77.44% involve only 1
query. Another finding was that the majority of visitors to the portal do not perform
any query [7]. Additionally the Max Planck Institute for Informatics analyzed the
verity server logs (action logs, user tracking) to research the user interaction behavior.
In particular, they focused on the query and result-click history. Concerning the
interface language selection, they found that the majority of users (84%) leaves the
User Behavior and Evaluation of Multilingual Information Access in Digital Libraries
default interface language English [7]. Another finding was that the most frequent
keywords relate to European place names or subjects.
Through a study of library catalog search logs, the CACAO Project6 found, that in
a library operating in a multicultural context, about 20% of the queries are written in
three languages, namely Italian, German and English [4].
The Europeana online survey conducted by the independent research agency IRN
Research determined that over 50% of all respondents or 69.6% of those reaching the
search results page refined the search by language [10].
The Eurovision system and the services of Tate Online were evaluated by
multilingual users [5] [14]. The key finding was that many users are more likely to
visit the collection site if it were translated into their preferred language and most of
the participants were willing to accept a text which they could understand but was not
perfectly translated. Within Multimatch7 two extensive user studies were organized
[15] [19]. The Clarity prototype was used to perform different tasks to explore
interaction issues. Among other things, they suggest that users should have the
possibility to choose the language they want to search in, depending on the individual
skills and the task.
The previous findings show a growing interest in MLIA issues but also
demonstrate that there are still a few open questions left. Especially inconsistent
statements require further research. For example have all studies in common, that
users are more likely to visit a Web site if it is translated into their preferred language.
However, the majority of users still leave the default English interface.
4 Proposed Research approach
4.1 Methodologies
I am going to use the combination of quantitative and qualitative research methods,
which provides different insight and therefore a more complete picture of users and
their behavior. Transaction log analysis (TLA) and in-depth usability tests as
complementary tools can allow a deeper understanding of users’ interaction with
information systems. The major advantage of TLA is that it automatically and
passively captures real users in their own daily environment. It is also an effective
way to detect discrepancies between what users say they do (for example in a survey
or interview) and what they actually do when they use an online system or web site
[6]. However, this method of analysis has a number of challenges and limitations. It is
nearly impossible to identify individual users with absolute accuracy. The same user
may use several IP addresses or several users can share one IP address. Using
hostnames to group or locate users geographically can also be misleading. Aside from
that, difficulties arise, when an attempt is made to answer questions concerning the
6 Maria Gäde
users´ motivations. Log entries are limited to the users’ interaction and do not reveal
backgrounds or preferences [23].
Questions that arise from the TLA and those which cannot be answered by it will
be addressed by the following usability test. The test design, including the tasks and
questions for the participants will be based on previous observations in general. Like
all research methods, there are a few limitations and pitfalls with qualitative user tests
as well. The obtained data cannot be generalized.
4.2 Effective Log file Analysis for Multilingual usage of search based web
sites – Europeana ClickStream Logger (CSL)
In the web, a transaction log is an electronic record of interactions during a search
session between the information system and the users searching for information [11].
Common log file entries are general and therefore contain limited information
concerning multilingual issues. Clickstream logging is a logging approach, which
enables to mine complex data in order to analyze user paths. The term "clickstream"
describes the path a user takes through a website. A clickstream is a series of actions
or requests on the web site accompanied by information on the activity being
performed [13]. It allows to track application state changes and therefore traces user
behavior in a way that a traditional http transaction log is unable to. For the
Europeana Clickstream Logs (CSL), different activity types or states with a particular
focus on multilingual access aspects are logged. Table shows an abbreviated log entry
for an interface language change:
Table 1. Abbreviated example from Europeana CSL with action: language change
["action" : "LANGUAGE_CHANGE", "agent" : "Mozilla/4.0
(compatible; MSIE 6.0; Windows NT 5.1; SV1; .NET CLR
1.1.4322; InfoPath.2; .NET CLR 2.0.50727)", "date" :
"2010-10-27T20:50:58.226+02:00", "invoked_at" : { "d" :
"2010-10-27", "t" : "20:50:58" }, "ip" : "ES", "isBot"
: false, "lang" : "ES", "oldLang" : "EN",…]
The log entry shows a user from Spain, who changes the interface language from the
default English to the Spanish translation. The URL of the requested page (where the
interface language could in this case also be reconstructed from), the referrer page, the
session and user id as well as page numbers are noted but not shown in this example.
Different user (search) patterns can be discovered which will be used to better
understand user behavior in a multilingual environment. Beside the general analysis
of user behavior the study will mainly focus on language information from the log
data such as interface language, country information from the IP address, query
language as well as language of the results viewed.
User Behavior and Evaluation of Multilingual Information Access in Digital Libraries
4.3 Understanding users` behavior
The second part of the analysis consists of usability tests which will answer questions
about the observed behaviors, dealing with the motivation and background of
different users. For Europeana seven Personas have been identified and characterized
according to their search behavior and literacy [8]. Currently the descriptions do not
contain any language information in order to be adaptable to all European Countries.
Through a usability test with subjects fitting to one Persona group from different
countries we want to determine whether different language skills and/or cultural
backgrounds influence their behavior.
Depending on the results from the usability test observations it could be necessary to
interview a small number of users from different countries to provide detailed context
data [3].
4.4 From User Requirements to System Requirements
The outcome of the analysis will be a catalogue of user requirements in a multilingual
environment. The challenge is to find a balance between user and system
requirements. Not everything users want is also feasible. Through the prioritization of
requirements, one possible design approach for interactive MLIA systems will be
Table 3 shows one example for the translation of user needs to system features.
Assumed users want to control the query translation by choosing different translation
candidates, the necessary conclusion would be an interactive MLIA system that
supports user assisted query translation, at least as an opportunity if things go wrong.
Table 2. Relation between User Behavior and System Requirements
User Requirement System Requirement
Transparent query translation Include user-assisted query translation
facilities. Support of translation
candidates/suggestions from the user
8 Maria Gäde
5 Summary
Figure 1 summarizes the outline of the research project as described above. Starting
from a state of the art overview of user studies in a multilingual environment a
combined research approach has been presented: Log file analysis supported by in-
depth usability tests. In the next few months the ClickStream Logs will be analyzed,
so it will be possible to compare data from a longer period of time. The interpretation
of the results will be used for the validation of the research question.
6 Discussion
I would especially benefit from presenting and discussing the Clickstream Logger
results, since this will influence the qualitative analysis significantly. It would be
helpful to receive feedback on the current log structure as well as the results and their
interpretation concerning multilingual issues. Furthermore the selection of
participants for the qualitative analysis and the criteria for prioritization of results
could be discussed with the mentors.
This research project is partly funded by EuropeanaConnect. Especially, I want to
thank Sjoerd Siebinga for developing the ClickStreamLogger.
1. Aula, A., Kellar, M.: Multilingual Search Strategies. In: CHI EA ’09: Proceedings of the
27th international conference extended abstracts on human factors in computing systems.
pp. 3865-3870 New York,: ACM (2009)
2. Bosca, A. and Dini, L.: CACAO Project at the LogCLEF Track. In: Working notes of the
Cross Language Evaluation Forum (CLEF) (Corfu, Greece, 30 September -2 October 2009)
3. Boyce, C., Neale, P.: Conducting In-Depth Interviews. A Guide for Designing and
Conducting In-Depth Interviews for Evaluation Input. (Pathfinder International Tool Series,
Monitoring and Evaluation – 2) (2006)
4. CACAO: D7.4 User Requirements for Advanced Features (2009)
5. Clough, P., Sanderson, M.: User Experiments with the Eurovision Cross-Language Image
Retrieval System. In: Journal of the American Society for Information Science and
Technology, 57(5), pp. 697 – 708 (2006)
6. Covey, D. T.: Usage and Usability Assessment. Library Practices and Concerns.
Washington, DC: Digital Library Federation (2002)
User Behavior and Evaluation of Multilingual Information Access in Digital Libraries
7. EDLproject: M1.4, Interim Report on Usability Developments in The European Library
8. EuropeanaConnect: Personas Catalogue V.2 (2010)
9. Hofmann, K., Rijke, M. de, B. Huurnink, B., Meij, E. J.: A Semantic Perspective on Query
Log Analysis. In: Working Notes for the CLEF 2009 Workshop (2009)
10. IRN Research: Europeana Online Visitor Survey Research Report Version 3 (2009)
11. Jansen, B. J.: Search log Analysis: What it is, what's been done, how to do it. In: Library &
Information Science Research, 28, pp. 407–432 (2006)
12. Janssen, Olaf: Gabriel 1997-2003 & Gabriel/TEL user survey (2003)
13. Joachims, T.: Optimizing Search Engines using Clickthrough Data. In: KDD ’02:
Proceedings of the eighth ACM SIGKDD international conference on Knowledge
discovery and data mining, pp. 133–142, New York, NY, USA, ACM Press (2002)
14. Minelli, S. H., Marlow, J., Clough, P., Cigarran Recuero, J.M., Gonzalo, J., Oomen, J. and
Loschiavo, D.: Gathering Requirements for Multilingual Search of Audiovisual Material in
Cultural Heritage. In: Proceedings of Workshop on User Centricity – state of the art (16th
IST Mobile and Wireless Communications Summit (2007)
15. MultiMatch: D1.2, User Requirements Analysis (2006)
16. Oakes, M., Xu, Y.: Search Log Analysis at the University of Sunderland. Paper presented
on the 10th Workshop of the Cross-Language Evaluation Forum (2009)
17. Oard, D.W.: Multilingual Information Access. In: Encyclopedia of Library and
Information Sciences, 3rd Ed. (2009)
18. Peters, C., Sheridan, P.: Multilingual Information Access. In: Agosti, M, Crestani, F, Pasi,
G (Eds.): Lectures on information RetrievalSpringer Lecture Notes In: Computer Science
Series, vol. 1980. Springer New York, New York, NY, pp. 51-80 (2001)
19. Petrelli, D., Beaulieu, M., Sanderson, M.: User Requirement Elicitation for Cross-language
Information Retrieval. In: The New Review of Information Behaviour Research, 3, pp. 17-
35 (2002)
20. Srinivasarao, V.: Mining the Behavior of Users in a Multilingual Information Access Task.
Cross Language Information Forum. In: Evaluation of Multilingual and Multi-modal
Information Retrieval: 9th Workshop of the Cross-Language Evaluation Forum
21. TELplus: D3.2, Improving Full-text Search in printed Digital Libraries’ Collections
through Semantic and Multilingual Functionalities - Technologies Assessment & User
Requirements (2009)
22. TELplus: D5.1, Report on User Requirements of the Target Library Services (2008)
23. Tenopir, C.: Use and Users of Electronic Library Resources: An Overview and Analysis of
Recent Research Studies (2003)
24. TrebleCLEF: D3.3, Best Practices in System-oriented and User-oriented Multilingual
Information Access (2009)
25. TrebleCLEF: D3.2, Workshop on Best Practices for the Development of Multilingual
Information Access Systems: the User Perspective (2008)
... Compared to system-centered studies, relatively few studies have focused on the user and aimed to explore users' behavior and expectations when interacting with Multilingual Digital Libraries. The majority of these studies employed mainly log analysis (Gäde, 2011;Ghorab et al., 2010), or a qualitative method which tends to generalize findings whereas a mixture of methods were employed by others Wu et al., 2010). The main goal of these studies was to shed light on users' information seeking behavior while searching in a multilingual environment and provide guidelines in designing interfaces. ...
... (continued ) Bilal & Bachir, 2007;Budzise-Weaver et al., 2012;Cheng et al., 1999;Clough & Eleta, 2010;Francis, 2008;Gäde, 2011;Kapidakis et al., 1999;Klavans, 1998;Nichols et al., 2005;Oard, 1997;Pavani, 2001;Pavlov et al., 2010;Petrelli & Clough, 2012;Ruecker et al., 2011;Sastry et al., 2011;Shiri, Ruecker, Doll et al., 2011;Stiller, 2011;Tripathi, 2008;Wang et al., 2006;Witten, 1997;Wu et al., 2012;Yang et al., 2000;Zeng, 2012 Bulletin of IEEE Technical Committee on Digital Libraries (2) (7), TUGboat Book section 8 Afifi, 2000;Bainbridge et al., 2003;Dartois et al., 1997;Jones et al., 2011;Monroy et al., 2007;Sheridan et al., 1997;Xu, 2003 Research and Advanced Technology for Digital Libraries (7), Knowledge-based information retrieval and filtering from the Web Report 9 Biagioni et al., 1998;CACAO, 2012;Europeana, 2009;Fox, 2000;Gey et al., 2006;Shivaram, 2002;XEROX, 2008 Cheng et al., 1999;Francis, 2008;Mizera-Pietraszko, 2009;Stiller, 2011;Stiller, Gäde & Petras, 2010;Singh, 2008;Wang, Teng et al., 2004;Wang et al., 2006;Yang et al., 2000;Zeng, 2012 1.2. Implementation and infrastructures of MLDLs 17 Afifi, 2000;Andreoni et al., 1999;Bhardwaj, 2010;CACAO, 2012;Fox, 2000;Kapidakis et al., 1999;Ruecker et al., 2011;Oard, 1997;Pavlov et al., 2010;Shiri et al., 2007;Shiri et al., 2010;Singh, 2008;8 E. Vassilakaki, E. Garoufallou ...
... Budzise Clinchant & Renders, 2009;Lee et al., 2003;Mizera-Pietraszko, 2009; Case study (5) Gäde, 2011;Wu et al., 2010;Wu et al., 2012 Log analysis, Interview (2), Questionnaire (3) 2.2. User behavior 4 Bilal & Bachir, 2007;Ghorab et al., 2010;Petrelli & Clough, 2012;Takaku et al., 2010 Log analysis (3) Shiri, Ruecker, Doll et al., 2011;Stafford et al., 2008 Interview ( Table 5, for each identified category and subcategories, the methods employed by the assigned papers are illustrated. ...
Full-text available
Purpose This study aims to identify, collect and critical review the research literature on Multilingual Digital Libraries in English language from 1997 to 2012. Design/methodology/approach The present literature review has followed the rules of systematic review. In particular, the identified relevant papers were categorized based on their expressed aim on two core themes, that of system-centered and user-centered studies. The assigned papers were further analyzed and six sub-themes emerged for the system-centered studies and four for the user-centered studies. Additional categorization was also provided according to type of publication. Findings The literature concerning Multilingual Digital Libraries is vast and mainly focuses on two aspects the “System” and the “Users”. The majority of papers tried to meet the challenges raised for enabling multilingual information retrieval in Digital Libraries. Unfortunately, these efforts undertaken by a small number of researchers or research groups apparently working in isolation and therefore resulting in the development of numerous different tools and techniques. Relatively few studies have focused on the user and aimed to explore users' behavior and expectations when interacting with Multilingual Digital Libraries. As a result, further research is needed to reach to some tangible and usable findings. Originality/value This literature review captures the diversity of the research conducted regarding multilingual information access and retrieval in Digital Libraries. It organizes the vast literature in comprehensive themes and sub-themes enabling easy access to specific information. Limitations This study reviews only papers in English due to language restrictions from 1997 to 2012.
... A. EXISTING MULTILINGUAL SUPPORT Font [5]Fonts have often been indiscriminately mapped to the same set of bytes e.g. 0x00 to 0xff are often used for both character and dingbats.But it have some drawbacks which are, there is requirement of installation on client machine. ...
... And there are conflicting national and industry standards because of use of multiple inconsistent character codes. GIST [5] GIST stands for Graphics and Intelligence Based Script Technology. It is solution for Indian languages which is hardware based and developed by C-DAC Still it has disadvantages, In this, character set is limited upto 256 values or characters because it uses 1 byte character representation.GIST can handle only one Indian language at a time It cannot be used without GIST card (hardware).It cannot be used for multi-lingual documents of Indian languages ...
Full-text available
This paper will focus on addressing the major challenges and issues in case of multilingualism aspects. The construction of multilingual web sites which is the best solution to addressing the problem occurred in the Internet facility of diverse cultural background. But still it causes some issue as like, to developing multiple instances of the same site in different languages causes increased overhead for the website implementation phase and also for the website maintenance phase. So probable solutions for these issues is use of UNICODE in ICT for India. By using UNICODE, design of multilingual editor for Web designing, HTML editor, scripting which have Indian language support. Also Adoption of open standards and open source software also investigates the benefits that are used in implementation of multilingual e-governance solutions.
... Maria Gäde's research supported these findings, noting that users are more likely to visit a page (including a search portal) if it is presented in their preferred language and positing that multilingualism can pose a major barrier to effective searching. 12 These user studies also agree that individuals who either do not speak English at all or who are not fluent have the most pressing needs for foreign language materials. Clough and Eleta noted that searchers who are not fluent may be able to understand a document written in English but be unable to construct the query necessary to retrieve it and suggested that cross-language searching would be most useful to these users. ...
Full-text available
As American society becomes more diverse, archivists increasingly work with multilingual collections and patrons. In Arizona, this situation occurs most frequently with materials created by individuals and communities using Spanish as their primary language. This case study discusses Arizona State University's creation of English and Spanish finding aids for six collections processed as part of a Council on Library and Information Resources grant. It describes the process of creating a Spanish finding aid template; reviews the challenges encountered and solutions designed while translating, encoding, and publishing Spanish guides; and analyzes use of the final documents.
Agroecology means that agriculture is a part of ecological systems. Agroecology thus promotes biodiversity and support multicultural production. Farmers are benefiting from the digital revolution that allow access to agroecological knowledge. Although internet access to information resources is becoming less problematic, the issue of language barrier is particularly critical. This chapter therefore focuses on the need for farmers to access useful information, with focus on language barriers. The linguistic issue is addressed using the Organic.Edunet experience ( Organic.Edunet is a learning portal that provides access to high-quality and trusted digital learning resources on organic agriculture and agroecology. These resources are used by students, teachers and farmers, as well as the general public interested in the subject. Organic.Edunet is used in this chapter as a use-case for analysing the benefit of truly multilingual portal in the agroecological field. Automated multilingual services introduced in the portal are described as well as the study of the analytics that shows the need to access information without the language barrier. A professional approach is described for demonstrating the benefit for farmers and teachers to use such thematic and multilingual portal. Then the importance of new content is mentioned to ensure the update of the information as well as the sustainability of such tools.
Full-text available
This paper presents the participation of the CACAO prototype to the Log Analysis for Digital Societies (LADS) task of LogCLEF 2009 track. CACAO (Cross-language Access to Catalogues And On-line libraries) is an EU project devoted to enabling cross-language access to the contents of a federation of digital libraries with a set of software tools for harvesting, indexing and serching over such data. In our experiment we investigated the possibility to exploit the TEL logs data as a source for inferring new translations, thus enriching already existing translation dictionaries; the proposed approach is based on the assumption that users consulting a multilingual digital col-lection are likely to repeat the same query in different languages. We applyed our approach to the logs from TEL and the results obtained are very promising.
This paper summarizes the participation of IIIT-H in the CLEF 2008 interactive task. Our goal was to mine the logs and extract conclusions about the behavior of users when facing a strictly multilingual information access task. We are provided the search logs which are generated by an online game, known-item image retrieval from Flickr. In this paper we describe the following tasks. We looked for the dierences in the search behavior according to the language skills. We clustered the users based on the score of the user, precision of the user and the number of hints he asked for. We then studied the behavior of the most successful user cluster, the least successful (unsuccessful) user cluster and the users in between the above two. Our results show that, most of the users start with monolingual interface and soon they realize cross-lingual is interface is more useful than mono-lingual interface, and the users are more comfortable to search in their mother language or the languages that they know.
We present our views on the CLEF log file analysis task. We argue for a task definition that focuses on the semantic enrichment of query logs. In addition, we discuss how additional information about the context in which queries are being made could further our understanding of users' information seeking and how to better facilitate this process.
The use of data stored in transaction logs of Web search engines, Intranets, and Web sites can provide valuable insight into understanding the information-searching process of online searchers. This understanding can enlighten information system design, interface development, and devising the information architecture for content collections. This article presents a review and foundation for conducting Web search transaction log analysis. A methodology is outlined consisting of three stages, which are collection, preparation, and analysis. The three stages of the methodology are presented in detail with discussions of goals, metrics, and processes at each stage. Critical terms in transaction log analysis for Web searching are defined. The strengths and limitations of transaction log analysis as a research method are presented. An application to log client-side interactions that supplements transaction logs is reported on, and the application is made available for use by the research community. Suggestions are provided on ways to leverage the strengths of, while addressing the limitations of, transaction log analysis for Web-searching research. Finally, a complete flat text transaction log from a commercial search engine is available as supplementary material with this manuscript.
Conference Paper
We explored the search strategies of multilingual searchers, i.e., users who use multiple languages when searching for information. We wanted to understand factors that determine the language multilingual searchers choose to search in, if they switch languages within a search task, and if they encounter challenges when searching in a non-native language. Our results indicate that availability and perceived quality of information were the primary reasons for searching in a non-native language. Language switching within a search only occurred when information could not be found with the original search language. We also observed a language-related use case where the goal was not to find information in a typical sense, but rather to check for correct phrases in the non-native language using search engines. Our research highlights several areas of future work for further understanding the multilingual search process.
Conference Paper
The global information society has radically changed the way in which know-ledge is acquired, disseminated and exchanged. Users of internationally distributed networks need to be able to find, retrieve and understand relevant information in whatever language and form it may have been stored. For this reason, much attention has been given over the past few years to the study and development of tools and technologies for multilingual information access (MLIA). This is a complex, multidisciplinary area in which methodologies and tools developed in the fields of information retrieval and natural language processing converge. Two main sectors are involved: multiple language recognition, manipulation and display; cross-language search and retrieval. The paper provides an overview of the main issues of interest in both these areas. Topics covered include: multilingual document indexing, specific requirements of particular languages and scripts, techniques for cross-language information retrieval (CLIR), resources, and system and component evaluation.
In this paper we present Eurovision, a text-based system for cross-language (CL) image retrieval. The system is evaluated by multilingual users for two search tasks with the system configured in English and five other languages. To our knowledge this is the first published set of user experiments for CL image retrieval. We show that: (1) it is possible to create a usable multilingual search engine using little knowledge of any language other than English, (2) categorizing images assists the user's search, and (3) there are differences in the way users search between the proposed search tasks. Based on the two search tasks and user feedback, we describe important aspects of any CL image retrieval system.