- Doctoral Forum
This year, 14 student presentations are accepted. The abstract of each presentation can be access by clicking on the title.Motivations for Image Tagging in Web 2.0 Applications Emma Angus, University of Wolverhampton, UK
This research project uses an information science and webometric approach to compare motivations to tag images in web 2.0 applications with tagging practices through the use of a questionnaire and tag classification analysis in order to find out what kinds of tags are created for images in different contexts, and why. This kind of information can help system designers to understand the contexts in which user-generated tags are likely to be helpful for information retrieval and perhaps also help them to redesign their systems to improve the chance that the tags entered are useful for retrieval purposes. The key result of the thesis will be a model of tagging practices that will encompass key factors affecting users (e.g., whether the tagged image is in a group, what type of image is being tagged, the age and gender of the tagger) and, based upon these, predict the most common types of tags to be created.
Analysis and retrieval of structured text Malcolm Clark, The Robert Gordon University, UK
This research focuses on overlaps of information retrieval (IR), cognitive science and genre, merging and utilizing these for one particular goal: to analyse and retrieve structured text. Structured textual documents are normally composed of several layers or sections which together form types, or genres, of text preserved, in particular, in e-mail and Wikipedia (XHTML). At present, the IR community of computer scientists and other researchers, such as computational linguists, is using genre to categorise documents in digitally structured media i.e. hypertext markup language (HTML) but more research needs to be done on linking cognitive science modelling techniques and IR. There has been much discussion on the definition of genre, but for the purpose of this project, genre is defined as purpose and form. This study intends to examine the attributes of genre to discover how they are used and, in particular, perceived.
Sensemaking of information during interactive information retrieval Abdigani Diriye, UCL Interaction Centre, UK
This research seeks to answer the question of how IR systems can support the user in sensemaking of information during interactive information retrieval In essence, this research deals with trying to enable the user to navigate within the information space more adeptly and effectively, so they are able to reach their information goal more efficiently. It is intended that work on query formulation will provide a greater capacity for the user to enter the information space and navigate within it more effectively. By improving the system's presentation of search results, the user is able to identify their location within the information space and formulate an accurate mental model of the space. Further work into document representations can provide clearer cues to identify the relevance of a document and whether endeavouring along this search path within the information space will satisfy their information need. These aspects should enable the user to search in a goal-oriented efficient manner within the information space.
Collaborative tagging as a user-generated knowledge organisation and validation tool Valerie Durieux, Université Libre de Bruxelles, Belgium
The objective of this research is to explore the ability of the collaborative tagging to equal or even surpass professional indexing – in other words: is collaborative tagging a reliable means to organise and validate online information? More specifically, this study focuses on the possibility to use collaborative tagging to improve information retrieval on the Web by organising and validating its content. The study consists of the following aspects:
- an overview of collaborative tagging, applications of tagging and practices of taggers
- evaluation and comparison of collaborative tagging and professional tagging in terms of quality
- the validation of online information.
Constructing, sharing and transferring health-related knowledge by using ICT. The context of preventive health care of obesity and metabolic syndrome. Heidi Enwald, University of Oulu, Finland
The growing popularity of the Internet has made it easier and faster to gain access to health information. Much of this information is valuable; however, the Internet also allows rapid and widespread distribution of false and misleading information. As the Internet is unregulated it has become important to evaluate the quality of information appearing online. Tailoring is one of the strategies that aim at enhancing message relevance. The more is known about the intended recipients of a communication, the better results are received from making the message relevant to them. The aim of this research is to shed new light on the documented content of health related knowledge that is shared and transferred by utilizing ICT.
Recruiting the public: Folksonomies and the potential usage as classification scheme in e-government context Huda R. Farhan, University of Sheffield, UK
This study is an attempt to fill in the gap in the literature regarding the quality of folksonomies as a classification tool at an e-government context. In addition, this study will develop evaluation criteria that can be used to evaluate folksonomies in any context. Furthermore, the importance of this study is that it tackles the little investigated topic of citizens’ information seeking experience at an e-government context. The main aims of this study are to investigate citizen’s satisfaction of the organization and classification of information at e-government web sites, in order to evaluate the efficiency of information retrieval in e-government web sites, and to propose classification methods that might improve information retrieval at e-government web sites through the investigation the value of folksonomies in the effectiveness of information retrieval process in the e-government context from citizens’ perspective.
Documentation and use of a client information system by social workers in child protection services Saila Huuskonen, University of Tampere, Finland
The aim of this research is to explore how social workers record and use information concerning clients as a part of their work and how they use client information system (CIS) in this process. The research has three major themes: essential information in child protection process, documentation and use of client information system. The focused research questions are as follows:
- What kind of information is recorded in different stages of protection process?
- What factors influence the selection of information to be recorded?
- What kind of information is considered essential when working with the case?
- For what purposes is information recorded?
- For what purposes is the recorded information used?
- How does the client information system support the documentation and use of documented information?
- How does the variation of institutional practises influence the documentation and use of the client information system?
- How does the variation of client information systems influence the documentation and use of the systems?
Dialogue Act Classification of Online Chat Reference Conversations for Information Retrieval Keisuke Inoue, Syracuse University, USA
The recent development in Web technologies has enabled an increase in computer-mediated information-seeking dialogues. The goal of the proposed study is to utilize such dialogues as a resource for information retrieval systems. In order to do so, this paper proposes a development of a Dialogue Act classification for online chat reference conversations. The contribution of the study will be to provide an information resource with rich specification that enables better information retrieval.
Tags in interactive information retrieval Yong-Mi Kim, University of Michigan, USA
This study examines the role that tags play in interactive information retrieval (IIR). In contrast to previous studies on tagging, the focus is on tag use, not the act of tagging. When reference is made to IR systems, both Web search engines such as Google, and Web sites where users can search or navigate the contents, such as Flickr are included. Little is known as yet about how tags assist the user in their interactions with an IR system – for example, what type of role tags play in relevance judgments, or in query reformulation. Therefore this study examines how users interact with tags as part of their interaction with an IR system. How taggers select the tags to attach to particular content, or methods to incorporate tags into searching or ranking algorithms, are out of the scope of this research.
Comment-Centric Blog Analysis Xiaozhong Liu, Syracuse University, USA
Blog analysis is becoming increasingly important in the context of Web2.0. The existing research includes blog retrieval, sentiment analysis, social network analysis and so on. This research proposes a new method of comment-centric blog analysis, namely, finding the reader-interested sub-topic(s) in the blog posting based on the information from reader’s comments to help us better understand the content of blog posting. Also proposed are approaches of experiments for automatically comments tagging, ranking and cross-media recommendation algorithms.
Tags as links for users to find new connections in a social network Charles Seger, Royal School of Library and Information Science, Denmark
This research project deals with the following issue: How can tags, indexed information objects or a combination of them function as a measure of connectivity between users in a folksonomy? This project explores the user context in a folksonomies by the use of informetric, network analytical and multivariable statistical methods. The aim is to investigate the collection of tags, user profiles and information objects in a case study of del.icio.us. The purpose is to collectively explore how users tag their resources, and how their collective indexing behaviour relates to certain fields of interest.
Automatic indexing methods for government documents Tanja Svarre, Royal School of Library and information Science, Denmark
The automatic indexing of governmental documents has not been examined to a large extent within the research literature. This project focuses on two issues: 1) gaining knowledge about the domain of e-government and the context-specific demands for indexing; and 2) examining how indexing methods perform in relation to the specific context and demands. Therefore, an identification and analysis of the specific characteristics of the domain of e-government is necessary e.g. of document genres, work tasks, information needs, information use and seeking behaviour. The purpose of this identification and analysis is to be able to plan and evaluate automatic indexing methods according to the information needs, and, in general, information use and seeking behaviour of public administration, specifically to electronic document management systems (EDMS). The main purpose of the project is to identify the most suitable approaches to automatic indexing when it comes to the characteristics of e-government.
Latin as pivot language for cross-language information retrieval Romance languages John Zacharias Theophanous, University of Pretoria, South Africa
This particular research study focuses upon the development of an algorithmic retrieval system, which may be utilised as a basis for the development and for the implementation of a dictionary-based, cross-language information retrieval (CLIR) system. This system utilises the Latin language as a pivot language for simultaneous retrieval in multiple Romance languages, with specific reference to Catalan, English, French, Italian, Portuguese, and Spanish. Several practical applications with respect to the proposed algorithmic retrieval model are demonstrated. Finally, conclusions are derived from this particular study, and recommendations for further research, in this particular field of cross-language information retrieval, are offered. The primary research question pertaining to this research study may be stated as follows: How would an algorithmic retrieval system be constructed and implanted for facilitating simultaneous retrieval in multiple Romance languages through the utilisation of Latin as a pivot language?
Motivation for Adopting Emergency Communication Technology in Community Settings: A Case Study Philip Fei Wu, University of Maryland, College Park, USA
The main purpose of this study is to understand various factors that affect community members’ motivation and ability to participate in ICT-enabled emergency response. The key questions my research intends to answer are: What motivates community members to adopt and use emergency response technologies? What prevents them from using such technologies? What can be done to lower the cognitive, social, and technical barriers of adopting emergency response systems? Answers to these questions not only will inform the system design and assist practitioners in deploying and promoting response systems, but will also provide useful insights to researchers interested in how technologies can facilitate communication and cooperation among community members, especially in response to high-stress, high-stakes situations.
Students will be given 15-20 minutes for the presentation of their research project. Each presentation will be followed by 15 to 20 minutes of discussion with senior researchers, which are experts in one or more of the general IIiX symposium themes:
Doctoral Forum Senior ResearchersKatriina Byström, Swedish School of Library and Information Science, Sweden
Theresa Dirndorfer-Anderson, University of Technology, Sydney, Australia
Ayse Goker, City University London, UK
Joemon Jose, University of Glasgow, UK
Diane Kelly, University of North Carolina, US
Daniella Petrelli, University of Sheffield, UK
LocationThe same as the symposium location (BCS London Office).
TimeOctober 14, 2008 between 9.00-17.00
ProgrammeThe doctoral forum programme can be accessed from here.