HCI Bibliography Home | HCI Conferences | UMAP Archive | Detailed Records | RefWorks | EndNote | Hide Abstracts
UMAP Tables of Contents: 00010203040506070809101112131415

Proceedings of User Modeling 2007 2007-07-25

Fullname:Proceedings of the 11th International Conference on User Modeling
Editors:Cristina Conati; Kathleen McCoy; Georgios Paliouras
Location:Corfu, Greece
Dates:2007-Jul-25 to 2007-Jul-29
Publisher:Springer-Verlag
Series:Lecture Notes in Computer Science, 2007, Volume 4511
Standard No:ISBN: 978-3-540-73077-4 (Print) 978-3-540-73078-1 (Online); hcibib: UMAP07
Papers:69
Pages:484
Links:Online Proceedings | Conference Home Page
Summary:As the variety and complexity of interactive systems increase, understanding how a system can dynamically capture relevant user needs and traits, and automatically adapting its interaction to this information, has become critical for devising effective advanced services and interfaces. The International User Modeling Conference represents the central forum for presenting the advances in the research and development of personalized, user-adaptive systems. Bi-annual scientific meetings of the user modeling community started in 1986 as a small invitational workshop held in Maria Laach, Germany, with 24 participants. The workshops continued with an open format, and grew into an international conference with 74 submissions in 1994. While maintaining its feel as a highly engaged and intimate community, the conference has continued to grow, reaching the record number of 169 submissions (153 full papers and 16 posters) in this current edition, held in Corfu, Greece.
    With an acceptance rate of 19.6% for long papers and 38% for posters, selected by a team of reviewers who proved to be exceptionally thorough and thoughtful in their reviewers, this year's program followed the high standards set by the previous editions, and presented an exciting range of interdisciplinary work covering topics such as cognitive modeling, modeling of user affect and meta-cognition, empirical evaluations of novel techniques, user modeling for mobile computing and recommender systems, user adaptivity and usability. In addition to 30 long paper presentations and 32 posters, this year's program featured 3 invited lectures, a doctoral consortium session with 5 student presentations, a demo program with 5 demos, 4 tutorials and 8 workshops. We continued the UM tradition of being a truly international event by having the first invited speaker from Asia (Yasuyoki Sumi from Japan), along with an invited speaker from North America (Martha Pollack from the USA) and one from Europe (Norbert Streitz from Germany). The international diversity was also reflected in the conference papers and posters with the geographical distribution of papers (posters) as follows: Europe 15 (14), Asia 3 (2), North America 10 (10), Australia/New Zealand 1 (3), Middle East 1 (2), South America 0 (1).
  1. Invited Papers
  2. Evaluating User/Student Modeling Techniques
  3. Data Mining and Machine Learning for User Modeling
  4. Collaborative Filtering and Recommender Systems
  5. Cognitive Modeling
  6. User Adaptation and Usability
  7. Modeling Affect and Meta-cognition
  8. Mobile, Ubiquitous, and Context Aware User Modeling
  9. Intelligent Information Retrieval, Information Filtering, and Content Personalization
  10. Poster Papers
  11. Doctoral Consortium Papers

Invited Papers

The Disappearing Computer: User-Centered Interaction Design for Smart Artefacts BIBAFull-Text 1-2
  Norbert Streitz
The increasing trend of embedding computation in everyday objects creating smart artefacts (Streitz et al., 2005 b) and the associated concept of the disappearing computer (Streitz, 2001, Streitz et al, 2007) raises new challenges for designing interactive systems. The unobtrusive character of this development is illustrated in this statement by Streitz and Nixon (2005): "It seems like a paradox but it will soon become reality: The rate at which computers disappear will be matched by the rate at which information technology will increasingly permeate our environment and our lives".
Experience Medium: Toward a New Medium for Exchanging Experiences BIBAFull-Text 3-4
  Yasuyuki Sumi
In this talk, I will propose a notion of "experience medium" in which we can exchange our experiences in museum touring, daily meetings, collaborative work, etc. The experience medium is a medium for capturing, interpreting, and creating our experiences, i.e., not only verbalized representations of our experiences but also their contextual information (awareness, common sense, atmosphere).
Intelligent Assistive Technology: The Present and the Future BIBAKFull-Text 5-6
  Martha E. Pollack
Recent advances in two areas of computer science -- wireless sensor networks and AI inference strategies -- have made it possible to envision a wide range of technologies that can improve the lives of people with physical, cognitive, and/or psycho-social impairments. To be effective, these systems must perform extensive user modeling in order to adapt to the changing needs and capabilities of their users. This invited talk provides a survey of current projects aimed at the development of intelligent assistive technology and describes further design challenges and opportunities.
Keywords: Assistive technology

Evaluating User/Student Modeling Techniques

Exploiting Evidence Analysis in Plan Recognition BIBAFull-Text 7-16
  Sandra Carberry; Stephanie Elzer
Information graphics, such as bar charts and line graphs, that appear in popular media generally have a message that they are intended to convey. We have developed a novel plan inference system that uses evidence in the form of communicative signals from the graphic to recognize the graphic designer's intended message. We contend that plan inference research would benefit from examining how each of its evidence sources impacts the system's success. This paper presents such an evidence analysis for the communicative signals that are captured in our plan inference system, and the paper shows how the results of this evidence analysis are informing our research on plan recognition and application systems.
Modeling the Acquisition of Fluent Skill in Educational Action Games BIBAFull-Text 17-26
  Ryan S. J. D. Baker; M. P. Jacob Habgood; Shaaron E. Ainsworth; Albert T. Corbett
There has been increasing interest in using games for education, but little investigation of how to model student learning within games [cf. 6]. We investigate how existing techniques for modeling the acquisition of fluent skill can be adapted to the context of an educational action game, Zombie Division. We discuss why this adaptation is necessarily different for educational action games than for other types of games, such as turn-based games. We demonstrate that gain in accuracy over time is straightforward to model using exponential learning curves, but that models of gain in speed over time must also take gameplay learning into account.
A User Modeling Server for Contemporary Adaptive Hypermedia: An Evaluation of the Push Approach to Evidence Propagation BIBAFull-Text 27-36
  Michael Yudelson; Peter Brusilovsky; Vladimir Zadorozhny
Despite the growing popularity of user modeling servers, little attention has been paid to optimizing and evaluating the performance of these servers. We argue that implementation issues and their influence on server performance should become the central focus of the user modeling community, since there is a sharply increasing real-life load on user modeling servers, This paper focuses on a specific implementation-level aspect of user modeling servers -- the choice of push or pull approaches to evidence propagation. We present a new push-based implementation of our user modeling server CUMULATE and compare its performance with the performance of the original pull-based CUMULATE server.

Data Mining and Machine Learning for User Modeling

Principles of Lifelong Learning for Predictive User Modeling BIBAFull-Text 37-46
  Ashish Kapoor; Eric Horvitz
Predictive user models often require a phase of effortful supervised training where cases are tagged with labels that represent the status of unobservable variables. We formulate and study principles of lifelong learning where training is ongoing over a prolonged period. In lifelong learning, decisions about extending a case library are made continuously by balancing the cost of acquiring values of hidden states with the long-term benefits of acquiring new labels. We highlight key principles by extending BusyBody, an application that learns to predict the cost of interrupting a user. We transform the prior BusyBody system into a lifelong learner and then review experiments that highlight the promise of the methods.
Users in Volatile Communities: Studying Active Participation and Community Evolution BIBAKFull-Text 47-56
  Tanja Falkowski; Myra Spiliopoulou
Active participation of a person in a community is a powerful indicator of the person's interests, preferences, beliefs and (often) social and demographic context. Community membership is part of a user's model and can contribute to tasks like personalized services, assistance and recommendations. However, a community member can be active or inactive. To what extend is a community still representative of the interests of an inactive participant? To gain insights to this question, we observe a community as an evolving social structure and study the effects of member fluctuation. We define a community as a high-level temporal structure composed of "community instances" that are defined conventionally through observable active participation and are captured at distinct timepoints. Thus, we capture community volatility, as evolution and discontinuation. This delivers us clues about the role of the community for its members, both for active and inactive ones. We have applied our model on a community exhibiting large fluctuation of members and acquired insights on the community-member interplay.
Keywords: user communities; community evolution; community participation
Learning from What Others Know: Privacy Preserving Cross System Personalization BIBAFull-Text 57-66
  Bhaskar Mehta
Recommender systems have been steadily gaining popularity and have been deployed by several service providers. Large scalable deployment has however highlighted one of the design problems of recommender systems: lack of interoperability. Users today often use multiple electronic systems offering recommendations, which cannot learn from one another. The result is that the end user has to often provide similar information and in some cases disjoint information. Intuitively, it seems that much can be improved with this situation: information learnt by one system could potentially be reused by another, to offer an overall improved personalization experience. In this paper, we provide an effective solution to this problem using Latent Semantic Models by learning a user model across multiple systems. A privacy preserving distributed framework is added around the traditional Probabilistic Latent Semantic Analysis framework, and practical aspects such as addition of new systems and items are also dealt with in this work.
Construction of Ontology-Based User Model for Web Personalization BIBAFull-Text 67-76
  Hui Zhang; Yu Song; Han-tao Song
Personalized Web browsing and search hope to provide Web information that matches a user's personal interests. A key feature in developing successful personalized Web applications is to build user model that accurately represents a user's interests. This paper deals with the problem of modeling Web users by means of personal ontology. A Web log preparation system discovering user's semantic navigation sessions is presented first. Such semantic sessions could be used as the input of constructing ontology-based user model. Our construction of user model is based on a semantic representation of the user activity. We build the user model without user interaction, automatically monitoring the user's browsing habits, constructing the user ontology from semantic sessions. Each semantic session updates the user model in such a way that the conceptual behavior history of the user is recorded in user ontology. After building the initial model from visited Web pages, techniques are investigated to estimate model convergence. In particular, the overall performance of our ontology-based user model is also presented and favorably compared to other model using a flat, unstructured list of topics in the experimental systems.

Collaborative Filtering and Recommender Systems

Preference-Based Organization Interfaces: Aiding User Critiques in Recommender Systems BIBAKFull-Text 77-86
  Li Chen; Pearl Pu
Users' critiques to the current recommendation form a crucial feedback mechanism for refining their preference models and improving a system's accuracy in recommendations that may better interest the user. In this paper, we present a novel approach to assist users in making critiques according to their stated and potentially hidden preferences. This approach is derived from our previous work on critique generation and organization techniques. Based on a collection of real user data, we conducted an experiment to compare our approach with three existing critique generation systems. Results show that our preference-based organization interface achieves the highest level of prediction accuracy in suggesting users' intended critiques and recommendation accuracy in locating users' target choices. In addition, it can potentially most efficiently save real users' interaction effort in decision making.
Keywords: Recommender systems; user preference models; critique generation; organization; decision support; experiment
"More Like This" or "Not for Me": Delivering Personalised Recommendations in Multi-user Environments BIBAKFull-Text 87-96
  David Bonnefoy; Makram Bouzid; Nicolas Lhuillier; Kevin Mercer
The television as a multi-user device presents some specificities with respect to personalisation. Recommendations should be provided both per-viewers as well as for a group. Recognising the inadequacy of traditional user modelling techniques with the constraint of television's lazy watching usage patterns, this paper presents a new recommendation mechanism based on anonymous user preferences and dynamic filtering of recommendations. Results from an initial user study indicate this mechanism was able to provide content recommendations to individual users within a multi-user environment with a high level of user satisfaction and without the need for user authentication or individual preference profile creation.
Keywords: Personalisation; recommendation; preference; user model; group
Feature-Weighted User Model for Recommender Systems BIBAFull-Text 97-106
  Panagiotis Symeonidis; Alexandros Nanopoulos; Yannis Manolopoulos
Recommender systems are gaining widespread acceptance in e-commerce applications to confront the "information overload" problem. Collaborative Filtering (CF) is a successful recommendation technique, which is based on past ratings of users with similar preferences. In contrast, Content-Based filtering (CB) assumes that each user operates independently. As a result, it exploits only information derived from document or item features. Both approaches have been extensively combined to improve the recommendation procedure. Most of these systems are hybrid: they run CF on the results of CB and vice versa. CF exploits information from the users and their ratings. CB exploits information from items and their features. In this paper, we construct a feature-weighted user profile to disclose the duality between users and features. Exploiting the correlation between users and features we reveal the real reasons of their rating behavior. We perform experimental comparison of the proposed method against the well-known CF, CB and a hybrid algorithm with a real data set. Our results show significant improvements, in terms of effectiveness.

Cognitive Modeling

Evaluating a Simulated Student Using Real Students Data for Training and Testing BIBAFull-Text 107-116
  Noboru Matsuda; William W. Cohen; Jonathan Sewall; Gustavo Lacerda
SimStudent is a machine-learning agent that learns cognitive skills by demonstration. It was originally developed as a building block of the Cognitive Tutor Authoring Tools (CTAT), so that the authors do not have to build a cognitive model by hand, but instead simply demonstrate solutions for SimStudent to automatically generate a cognitive model. The SimStudent technology could then be used to model human students' performance as well. To evaluate the applicability of SimStudent as a tool for modeling real students, we applied SimStudent to a genuine learning log gathered from classroom experiments with the Algebra I Cognitive Tutor. Such data can be seen as the human students' "demonstrations" of how to solve problems. The results from an empirical study show that SimStudent can indeed model human students' performance. After training on 20 problems solved by a group of human students, a cognitive model generated by SimStudent explained 82% of the problem-solving steps performed correctly by another group of human students.
Modeling Students' Natural Language Explanations BIBAFull-Text 117-126
  Albert Corbett; Angela Wagner; Sharon Lesgold; Harry Ulrich; Scott Stevens
Intelligent tutoring systems have achieved demonstrable success in supporting formal problem solving. More recently such systems have begun incorporating student explanations of problem solutions. Typically, these natural language explanations are entered with menus, but some ITSs accept open-ended typed inputs. Typed inputs require more work by both developers and students and evaluations of the added value for learning outcomes has been mixed. This paper examines whether typed input can yield more accurate student modeling than menu-based input. This paper examines the application of Knowledge Tracing student modeling to natural language inputs and examines the standard Knowledge Tracing definition of errors. The analyses indicate that typed explanations can yield more predictive models of student test performance than menu-based explanations and that focusing on semantic errors can further improve predictive accuracy.
Applications for Cognitive User Modeling BIBAKFull-Text 127-136
  Marcus Heinath; Jeronimo Dzaack; Andre Wiesner; Leon Urbas
Usability of complex dynamic human computer interfaces can be evaluated by cognitive modeling to investigate cognitive processes and their underlying structures. Even though the prediction of human behavior can help to detect errors in the interaction design and cognitive demands of the future user the method is not widely applied. The time-consuming transformation of a problem "in the world" into a "computational model" and the lack of fine-grained simulation data analysis are mainly responsible for this. Having realized these drawbacks we developed HTAmap and SimTrA to simplify the development and analysis of cognitive models. HTAmap, a high-level framework for cognitive modeling, aims to reduce the modeling effort. SimTrA supports the analysis of cognitive model data on an overall and microstructure level and enables the comparison of simulated data with empirical data. This paper describes both concepts and shows their practicability on an example in the domain of process control.
Keywords: usability evaluation; human computer interaction; cognitive modeling; high-level description; analysis
Identifiability: A Fundamental Problem of Student Modeling BIBAFull-Text 137-146
  Joseph E. Beck; Kai-min Chang
In this paper we show how model identifiability is an issue for student modeling: observed student performance corresponds to an infinite family of possible model parameter estimates, all of which make identical predictions about student performance. However, these parameter estimates make different claims, some of which are clearly incorrect, about the student's unobservable internal knowledge. We propose methods for evaluating these models to find ones that are more plausible. Specifically, we present an approach using Dirichlet priors to bias model search that results in a statistically reliable improvement in predictive accuracy (AUC of 0.620 ± 0.002 vs. 0.614 ± 0.002). Furthermore, the parameters associated with this model provide more plausible estimates of student learning, and better track with known properties of students' background knowledge. The main conclusion is that prior beliefs are necessary to bias the student modeling search, and even large quantities of performance data alone are insufficient to properly estimate the model.

User Adaptation and Usability

Understanding the Utility of Rationale in a Mixed-Initiative System for GUI Customization BIBAFull-Text 147-156
  Andrea Bunt; Joanna McGrenere; Cristina Conati
In this paper, we investigate the utility of providing users with the system's rationale in a mixed-initiative system for GUI customization. An evaluation comparing a version of the system with and without the rationale suggested that rationale is wanted by many users, leading to increased trust, understandability and predictability, but that not all users want or need the information.
Respecting Users' Individual Privacy Constraints in Web Personalization BIBAFull-Text 157-166
  Yang Wang; Alfred Kobsa
Web personalization has demonstrated to be advantageous for both online customers and vendors. However, its benefits may be severely counter acted by privacy constraints. Personalized systems need to take users' privacy concerns into account, as well as privacy laws and industry self-regulation that may be in effect. In this paper, we first discuss how these constraints may affect web-based personalized systems. We then explain in what way current approaches to this problem fall short of their aims, specifically regarding the need to tailor privacy to the constraints of each individual user. We present a dynamic privacy-enhancing user modeling framework as a superior alternative, which is based on a software product line architecture. Our system dynamically selects personalization methods during runtime that respect users' current privacy concerns as well as the privacy laws and regulations that apply to them.
Personalized Previews of Alternative Routes in Virtual Environments BIBAKFull-Text 167-176
  Mehmed Kantardzic; Pedram Sadeghian
In virtual environments (VEs) there are often many routes available to a destination and each route has its own unique characteristics. At the same time, each VE user has unique preferences when selecting a route. In this paper, we present a new methodology for developing a personalized route preview interface, which provides a preview of several routes to a destination that closely match the user's preferences. The user's preferences are modeled by an automatically generated user profile that does not require any explicit input from the user. Simulation experiments have shown the potential of this approach in both accurately learning the user's preferences and personalizing the preview interface.
Keywords: Virtual Environments; Route Selection; Previews; Personalization
Visual Attention in Open Learner Model Presentations: An Eye-Tracking Investigation BIBAFull-Text 177-186
  Susan Bull; Neil Cooke; Andrew Mabbott
Using an eye-tracker, this paper investigates the information that learners visually attend to in their open learner model, and the degree to which this is related to the method of displaying the model to the learner. Participants were fourteen final year undergraduate students using six views of their learner model data. Results suggest some views of the learner model information may be more likely to encourage learners to inspect information about their level of knowledge, whereas in other views attention is directed more towards scanning the view, resulting in a lower proportion of time focussed on knowledge-related data. In some views there was a difference according to whether the learner model view was one of the participants' preferred formats for accessing their learner model information, while in other views there was little difference. This has implications for the design of open learner model views in systems opening the learner model to the learner for different purposes.

Modeling Affect and Meta-cognition

EEG-Related Changes in Cognitive Workload, Engagement and Distraction as Students Acquire Problem Solving Skills BIBAFull-Text 187-196
  Ronald H. Stevens; Trysha Galloway; Chris Berka
We have begun to model changes in electroencephalography (EEG)-derived measures of cognitive workload, engagement and distraction as individuals developed and refined their problem solving skills in science. For the same problem solving scenario(s) there were significant differences in the levels and dynamics of these three metrics. As expected, workload increased when students were presented with problem sets of greater difficulty. Less expected, however, was the finding that as skills increased, the levels of workload did not decrease accordingly. When these indices were measured across the navigation, decision, and display events within the simulations significant differences in workload and engagement were often observed. Similarly, event-related differences in these categories across a series of the tasks were also often observed, but were highly variable across individuals.
Eliciting Motivation Knowledge from Log Files Towards Motivation Diagnosis for Adaptive Systems BIBAKFull-Text 197-206
  Mihaela Cocea; Stephan Weibelzahl
Motivation is well-known for its importance in learning and its influence on cognitive processes. Adaptive systems would greatly benefit from having a user model of the learner's motivation, especially if integrated with information about knowledge. In this paper a log file analysis for eliciting motivation knowledge is presented, as a first step towards a user model for motivation. Several data mining techniques are used in order to find the best method and the best indicators for disengagement prediction. Results show a very good level of prediction: around 87% correctly predicted instances of all three levels of engagement and 93% correctly predicted instances of disengagement. Data sets with reduced attribute sets show similar results, indicating that engagement level can be predicted from information like reading pages and taking tests, which are common to most e-Learning systems.
Keywords: e-Learning; motivation; log files analysis; data mining; adaptive systems; user modeling
Assessing Learner's Scientific Inquiry Skills Across Time: A Dynamic Bayesian Network Approach BIBAKFull-Text 207-216
  Choo-Yee Ting; Mohammad Reza Beik Zadeh
In this article, we develop and evaluate three Dynamic Bayesian Network (DBN) models for assessing temporally variable learner scientific inquiry skills (Hypothesis Generation and Variable Identification) in INQPRO learning environment. Empirical studies were carried out to examine the matching accuracies and identify the models' drawbacks. We demonstrate how the insights gained from a preceding model have eventually led to the improvement of subsequent models. In this study, the entire evaluation process involved 6 domain experts and 61 human learners. The matching accuracies of the models are measured by (1) comparing with the results gathered from the pretest, posttest, and learner's self-rating scores; and (2) comments given by domain experts based on learners' interaction logs and the graph patterns exhibited by the models.
Keywords: Scientific Inquiry Skills; Dynamic Bayesian Networks
From Modelling Domain Knowledge to Metacognitive Skills: Extending a Constraint-Based Tutoring System to Support Collaboration BIBAFull-Text 217-227
  Nilufar Baghaei; Antonija Mitrovic
Constraint-based tutors have been shown to increase individual learning in real classroom studies, but would become even more effective if they provided support for collaboration. COLLECT-UML is a constraint-based intelligent tutoring system that teaches object-oriented analysis and design using Unified Modelling Language. Being one of constraint-based tutors, COLLECT-UML represents the domain knowledge as a set of constraints. However, it is the first system to also represent a higher-level skill such as collaboration using the same formalism. We started by developing a single-user ITS. The system was evaluated in a real classroom, and the results showed that students' performance increased significantly. In this paper, we present our experiences in extending the system to provide support for collaboration as well as problem-solving. The effectiveness of the system was evaluated in a study conducted at the University of Canterbury in May 2006. In addition to improved problem-solving skills, the participants both acquired declarative knowledge about good collaboration and did collaborate more effectively. The results, therefore, show that Constraint-Based Modelling is an effective technique for modelling and supporting collaboration skills.

Mobile, Ubiquitous, and Context Aware User Modeling

Mobile Opportunistic Planning: Methods and Models BIBAFull-Text 228-237
  Eric Horvitz; Paul Koch; Muru Subramani
We present a study exploring the promise of developing computational systems to support the discovery and execution of opportunistic activities in mobile settings. We introduce the challenge of mobile opportunistic planning, describe a prototype named Mobile Commodities, and focus on the construction and use of probabilistic user models to infer the cost of time required to execute opportunistic plans.
Analyzing Museum Visitors' Behavior Patterns BIBAFull-Text 238-246
  Massimo Zancanaro; Tsvi Kuflik; Zvi Boger; Dina Goren-Bar; Dan Goldwasser
Many studies have investigated personalized information presentation in the context of mobile museum guides. In order to provide such a service, information about museum visitors has to be collected and visitors have to be monitored and modelled in a non-intrusive manner. This can be done by using known museum visiting styles to classify the visiting style of visitors as they start their visit. Past research applied ethnographic observations of the behaviour of visitors and qualitative analysis (mainly site studies and interviews with staff) in several museums to define visiting styles. The current work validates past ethnographic research by applying unsupervised learning approaches to visitors classification. By providing quantitative empirical evidence for a qualitative theory we claim that, from the point of view of assessing the suitability of a qualitative theory in a given scenario, this approach is as valid as a manual annotation of museum visiting styles.
A Context-Aware Movie Preference Model Using a Bayesian Network for Recommendation and Promotion BIBAFull-Text 247-257
  Chihiro Ono; Mori Kurokawa; Yoichi Motomura; Hideki Asoh
This paper proposes a novel approach for constructing users' movie preference models using Bayesian networks. The advantages of the constructed preference models are 1) consideration of users' context in addition to users' personality, 2) multiple applications, such as recommendation and promotion. Data acquisition process through a WWW questionnaire survey and a Bayesian network model construction process using the data are described. The effectiveness of the constructed model in terms of recommendation and promotion is also demonstrated through experiments.
Intrinsic Motivational Factors for the Intention to Use Adaptive Technology: Validation of a Causal Model BIBAKFull-Text 258-267
  Fabio Pianesi; Ilenia Graziola; Massimo Zancanaro
In this paper, we propose and validate a model of the 'intention to use' adaptive audio-video guides in a museum setting, extending TAM to include intrinsic motivational factors (involvement, attention) and constructs specific to adaptivity (control, personalization). The results of a PLS analysis ran on the data from 115 subjects show that for adaptive museum guides intention to use is not affected by such traditional construct as perceived ease of use, whereas perceived usefulness and enjoyment play an important role. Also, both personalization and control are causally relevant, the former affecting enjoyment and the latter the perceived usefulness.
Keywords: Causal modelling; intention to use; adaptive systems; technology acceptance model; intrinsic motivations; structural equation modelling

Intelligent Information Retrieval, Information Filtering, and Content Personalization

Improving Social Filtering Techniques Through WordNet-Based User Profiles BIBAFull-Text 268-277
  Pasquale Lops; Marco Degemmis; Giovanni Semeraro
Collaborative filtering algorithms predict the preferences of a user for an item by weighting the contributions of similar users, called neighbors, for that item. Similarity between users is computed by comparing their rating styles, i.e. the set of ratings given on the same items. Unfortunately, similarity between users is computable only if they have common rated items. The main contribution of this paper is a (content-collaborative) hybrid recommender system which overcomes this limitation by computing similarity between users on the ground of their content-based profiles. Traditional keyword-based profiles are unable to capture the semantics of user interests, due to the natural language ambiguity. A distinctive feature of the proposed technique is that a statistical model of the user interests is obtained by machine learning techniques integrated with linguistic knowledge contained in the WordNet lexical database. This model, called the semantic user profile, is exploited by the hybrid recommender in the neighborhood formation process. The results of an experimental session in a movie recommendation scenario demonstrate the effectiveness of the proposed approach.
Push-Poll Recommender System: Supporting Word of Mouth BIBAFull-Text 278-287
  Andrew Webster; Julita Vassileva
Recommender systems produce social networks as a side effect of predicting what users will like. However, the potential for these social networks to aid in recommending items is largely ignored. We propose a recommender system that works directly with these networks to distribute and recommend items: the informal exchange of information (word of mouth communication) is supported rather than replaced. The paper describes the push-poll approach and evaluates its performance at predicting user ratings for movies against a collaborative filtering algorithm. Overall, the push-poll approach performs significantly better while being computationally efficient and suitable for dynamic domains (e.g. recommending items from RSS feeds).
Evaluation of Modeling Music Similarity Perception Via Feature Subset Selection BIBAFull-Text 288-297
  D. N. Sotiropoulos; A. S. Lampropoulos; G. A. Tsihrintzis
In this paper, we describe and discuss the evaluation process and results of a content-based music retrieval system that we have developed. In our system, user models embody the ability of evolving and using different music similarity measures for different users. Specifically, a user-supplied relevance feedback and related neural network-based incremental learning procedures allows our system to determine which subset of a set of objective acoustic features approximates more efficiently the subjective music similarity perception of an individual user. The evaluation results verify our hypothesis of a direct relation between subjective music similarity perception and objective acoustic feature subsets. Moreover, it is shown that, after training, retrieved music pieces exhibit significantly improved perceived similarity to user-targeted music pieces.

Poster Papers

A Practical Activity Capture Framework for Personal, Lifetime User Modeling BIBAFull-Text 298-302
  Max Van Kleek; Howard E. Shrobe
This paper addresses the problem of capturing rich, long-term personal activity logs of users' interactions with their workstations, for the purpose of deriving predictive, personal user models. Our architecture addresses a number of practical problems with activity capture, including incorporating heterogeneous information from different applications, measuring phenomena with different rates of change, efficiently scheduling knowledge sources, incrementally evolving knowledge representations, and incorporating prior knowledge to combine low-level observations into interpretations better suited for user modeling tasks. We demonstrate that the computational and memory demands of general activity capture are well within reasonable limits even on today's hardware and software platforms.
A Probabilistic Relational Student Model for Virtual Laboratories BIBAFull-Text 303-308
  Julieta Noguez; L. Enrique Sucar; Enrique Espinosa
We have developed a novel student model based on probabilistic relational models (PRMs). This model combines the advantages of Bayesian networks and object-oriented systems. It facilitates knowledge acquisition and makes it easier to apply the model for different domains. The model is oriented towards virtual laboratories, in which a student interacts by doing experiments in a simulated or remote environment. It represents the students' knowledge at different levels of granularity, combining the performance and exploration behavior in several experiments, to decide the best way to guide the student in the next experiments. Based on this model, we have developed tutors for virtual laboratories in different domains. An evaluation of with a group of students, show a significant improvement in learning when a tutor based on the PRM model is incorporated to a virtual robotics lab.
A Semantics-Based Dialogue for Interoperability of User-Adaptive Systems in a Ubiquitous Environment BIBAFull-Text 309-313
  Federica Cena; Lora M. Aroyo
In this paper we present an approach to enable interoperability of user-adaptive systems (UASs) in a ubiquitous environment. We model the interactions between systems as a semantics-based dialogue for exchanging user model and context data. We focus on the user data clarification and negotiation tasks, and show how semantics enables, on the one hand, the understanding among user-adaptive systems in a distributed ubiquitous setting, and on the other hand indirectly improves their effectiveness in producing end user results. We deploy and evaluate our approach in the UbiquiTO mobile adaptive tourist guide.
A User Independent, Biosignal Based, Emotion Recognition Method BIBAKFull-Text 314-318
  G. Rigas; C. D. Katsis; G. Ganiatsas; D. I. Fotiadis
A physiological signal based emotion recognition method, for the assessment of three emotional classes: happiness, disgust and fear, is presented. Our approach consists of four steps: (i) biosignal acquisition, (ii) biosignal preprocessing and feature extraction, (iii) feature selection and (iv) classification. The input signals are facial electromyograms, the electrocardiogram, the respiration and the electrodermal skin response. We have constructed a dataset which consists of 9 healthy subjects. Moreover we present preliminary results which indicate on average, accuracy rates of 0.48,0.68 and 0.69 for recognition of happiness, disgust and fear emotions, respectively.
Keywords: emotion recognition; biosignals; classification
A User Model of Psycho-physiological Measure of Emotion BIBAFull-Text 319-323
  Olivier Villon; Christine Lisetti
The interpretation of physiological signals in terms of emotion requires an appropriate mapping between physiological features and emotion representations. We present a user model associating psychological and physiological representation of emotion in order to bring findings from the psychophysiology domain into User-Modeling computational techniques. We discuss results based on an experiment we performed based on bio-sensors to get physiological measure of emotion, involving 40 subjects.
A User-Item Predictive Model for Collaborative Filtering Recommendation BIBAFull-Text 324-328
  Heung-Nam Kim; Ae-Ttie Ji; Cheol Yeon; Geun-Sik Jo
Collaborative Filtering recommender systems, one of the most representative systems for personalized recommendations in E-commerce, enable users to find the useful information easily. But traditional CF suffers from some weaknesses: scalability and real-time performance. To address these issues, we present a novel model-based CF approach to provide efficient recommendations. In addition, we propose a new method of building a model with dynamic updates, when users present explicit feedback. The experimental evaluation on MovieLens datasets shows that our method offers reasonable prediction quality as good as the best of user-based Pearson correlation coefficient algorithm.
Automatic Generation of Students' Conceptual Models from Answers in Plain Text BIBAFull-Text 329-333
  D. Pérez-Marín; E. Alfonseca; P. Rodríguez; I. Pascual-Nieto
Recently, we have introduced a new procedure to automatically generate students' conceptual models to assist teachers in finding out their students' main misconceptions and lack of concepts, from their interaction with an automatic and adaptive free-text scoring system. In this paper, we present an improvement of this procedure: the models can be built from the students' answers in plain text and they refer not only to one particular student but to the whole class. We also introduce a new tool called COMOV (COnceptual MOdels Viewer) to display the models as concept maps, tables, bar charts or text summaries. Finally, we provide an evaluation of this new approach.
Capturing User Interests by Both Exploitation and Exploration BIBAFull-Text 334-339
  Ka Cheung Sia; Shenghuo Zhu; Yun Chi; Koji Hino; Belle L. Tseng
Personalization is one of the important research issues in the areas of information retrieval and Web search. Providing personalized services that are tailored toward the specific preferences and interests of a given user can enhance her experience and satisfaction. However, to effectively capture user interests is a challenging research problem. Some challenges include how to quickly capture user interests in an unobtrusive way, how to provide diversified recommendations, and how to track the drifts of user interests in a timely fashion. In this paper, we propose a model for learning user interests and an algorithm that actively captures user interests through an interactive recommendation process. The key advantage of our algorithm is that it takes into account both exploitation (recommending items that belong to users' core interest) and exploration (discovering potential interests of users). Extensive experiments using synthetic data and a user study show that our algorithm can quickly capture diversified user interests in an unobtrusive way, even when the user interests may drift along time.
Conceptualizing Student Models for ICALL BIBAFull-Text 340-344
  Luiz Amaral; Detmar Meurers
Student models for Intelligent Computer Assisted Language Learning (ICALL) have largely focused on the acquisition of grammatical structures. In this paper, we motivate a broader perspective of student models for ICALL that incorporates insights from current research on second language acquisition and language testing. We argue for a student model that includes a representation of the learner's ability to use language to perform tasks as well as an explicit activity model that provides information on the language tasks and the inferences for the student model they support.
Context-Dependent User Modelling for Smart Homes BIBAKFull-Text 345-349
  Elena Vildjiounaite; Otilia Kocsis; Vesa Kyllönen; Basilis Kladis
This works presents a user modelling service for a Smart Home -- intelligent context-aware environment, providing personalized proactive support to its inhabitants. Diversity of Smart Home applications imposes various technical and implementation requirements, such as the need to model dependency of user preferences on context in a unified and convenient way, both for users and for application developers. This paper introduces the service architecture and currently implemented functionalities: stereotypes-based profiles initialisation; a GUI for acquisition of context-dependent and context-independent preferences, which provides an easy way to create own concepts of context ontology and to map them into already existing concepts; and a method to learn context-dependent user preferences from interaction history.
Keywords: User Model; Context Awareness; Smart Home
Conversations Amidst Computing: A Study of Interruptions and Recovery of Task Activity BIBAKFull-Text 350-354
  Shamsi T. Iqbal; Eric Horvitz
We present results from a field study investigating the influence of conversations on the multitasking behavior of computer users. We report on several findings, including the timing of the resumption of tasks following conversational interruptions and on the nature and rate of computing activities that are performed concurrently with conversation.
Keywords: Interruption; disruption; recovery; conversation; cognitive models
Cross-Domain Mediation in Collaborative Filtering BIBAFull-Text 355-359
  Shlomo Berkovsky; Tsvi Kuflik; Francesco Ricci
One of the main problems of collaborative filtering recommenders is the sparsity of the ratings in the users-items matrix, and its negative effect on the prediction accuracy. This paper addresses this issue applying cross-domain mediation of collaborative user models, i.e., importing and aggregating vectors of users' ratings stored by collaborative systems operating in different application domains. The paper presents several mediation approaches and initial experimental evaluation demonstrating that the mediation can improve the accuracy of the generated predictions.
Driver Destination Models BIBAKFull-Text 360-364
  John Krumm; Eric Horvitz
Predictive models of destinations represent an opportunity in the context of the increasing availability and sophistication of in-car driving aids. We present analyses of drivers' destinations based on GPS data recorded from 180 volunteer subjects. We focus on the probability of observing drivers visit previously unobserved destinations given time of day and day of week, and the rate of decline of observing such new destinations with time. For the latter, we discover a statistically significant difference based on gender.
Keywords: driving; mobility; destinations; cars; automobiles; navigation
Enabling Efficient Real Time User Modeling in On-Line Campus BIBAFull-Text 365-369
  Santi Caballé; Fatos Xhafa; Thanasis Daradoumis; Raul Fernandez
User modelling in on-line distance learning is an important research field focusing on two important aspects: describing and predicting students' actions and intentions as well as adapting the learning process to students' features, habits, interests, preferences, and so on. The aim is to greatly stimulate and improve the learning experience. In this context, user modeling implies a constant processing and analysis of user interaction data during long-term learning activities, which produces large and considerably complex information. As a consequence, processing this information is costly and requires computational capacity beyond that of a single computer. In order to overcome this obstacle, in this paper we show how a parallel processing approach can considerably decrease the time of processing log data that come from on-line distance educational web-based systems. The results of our study show the feasibility of using Grid middleware to speed and scale up the processing of log data and thus achieving an efficient and dynamic user modeling in on-line distance learning.
Eliciting Adaptation Knowledge from On-Line Tutors to Increase Motivation BIBAFull-Text 370-374
  Teresa Hurley; Stephan Weibelzahl
In the classroom, teachers know how to motivate their students and how to exploit this knowledge to adapt or optimize their instruction when a student shows signs of demotivation. In on-line learning environments it is much more difficult to assess a learner's motivation and to have adaptive intervention strategies and rules of application to help prevent attrition or drop-out. In this paper, we present results from a survey of on-line tutors on how they motivate their learners. These results will inform the development of an adaptation engine by extracting and validating selection rules for strategies to increase motivation depending on the learner's self-efficacy, goal orientation, locus of control and perceived task difficulty in adaptive Intelligent Tutoring Systems.
Improving User Taught Task Models BIBAFull-Text 375-379
  Phillip Michalak; James Allen
Task models are essential components in many approaches to user modelling because they provide the context with which to interpret, predict, and respond to user behavior. The quality of such models is critical to their ability to support these functions. This paper describes work on improving task models that are automatically acquired from demonstration. Modifications to a standard planning algorithm are described and applied to an example learned task model, showing the utility of incorporating plan-based reasoning into task learning systems.
Inducing User Affect Recognition Models for Task-Oriented Environments BIBAFull-Text 380-384
  Sunyoung Lee; Scott W. McQuiggan; James C. Lester
Accurately recognizing users' affective states could contribute to more productive and enjoyable interactions, particularly for task-oriented learning environments. In addition to using physiological data, affect recognition models can leverage knowledge of task structure and user goals to effectively reason about users' affective states. In this paper we present an inductive approach to recognizing users' affective states based on appraisal theory, a motivational-affect account of cognition in which individuals' emotions are generated in response to their assessment of how their actions and events in the environment relate to their goals. Rather than manually creating the models, the models are learned from training sessions in which (1) physiological data, (2) information about users' goals and actions, and (3) environmental information are recorded from traces produced by users performing a range of tasks in a virtual environment. An empirical evaluation with a task-oriented learning environment testbed suggests that an inductive approach can learn accurate models and that appraisal-based models exploiting knowledge of task structure and user goals can outperform purely physiologically-based models.
Interactive User Modeling for Personalized Access to Museum Collections: The Rijksmuseum Case Study BIBAKFull-Text 385-389
  Yiwen Wang; Lora M. Aroyo; Natalia Stash; Lloyd Rutledge
In this paper we present an approach for personalized access to museum collections. We use a RDF/OWL specification of the Rijksmuseum Amsterdam collections as a driver for an interactive dialog. The user gives his/her judgment on the artefacts, indicating likes or dislikes. The elicited user model is further used for generating recommendations of artefacts and topics. In this way we support exploration and discovery of information in museum collections. A user study provided insights in characteristics of our target user group, and showed how novice and expert users employ their background knowledge and implicit interest in order to elicit their art preference in the museum collections.
Keywords: CHIP (Cultural Heritage Information Presentation); user study; adaptive system; personalization; RDF/OWL; recommendations; user modeling
Kansei Processing Agent for Personalizing Retrieval BIBAKFull-Text 390-394
  Sunkyoung Baek; Myunggwon Hwang; Pankoo Kim
In the present, methods of creating and processing a profile are insufficient for achieving personalization information retrieval that reflects the subjective Kansei preference of users. To rectify this insufficiency, we have created a Kansei information processing agent. Our study proposes a Kansei agent for the creation, accumulation and renewal of profiles in personalized retrieval and explores possible contributions to the development of a Kansei-based recommendation system and personalizing service.
Keywords: Kansei; Kansei Processing; Personalizing Retrieval
Maximizing the Utility of Situated Public Displays BIBAFull-Text 395-399
  Jörg Müller; Antonio Krüger; Tsvi Kuflik
Situated public displays are intended to convey important information to a large and heterogeneous population. Because of the heterogeneity of the population, they may risk providing a lot of irrelevant information. Many such important information items presented on public displays are actionables, items that are intended to trigger specific actions. The expected utility that such actionables have for a user depend on the value of the action for the user. A goal should be to provide for each user the actionables with highest utility. This can be achieved by adapting the information presentation to the users currently in front of the display. Adaptation can take place either by identifying individual users, by using statistics about the user groups usually in front of the display or by a combination of both. We present a formal framework based on decision theory that enables the integration of sensor data and statistics and allows to choose the optimal actionable to present based on this data.
Modeling Preferences in a Distributed Recommender System BIBAFull-Text 400-404
  Sylvain Castagnos; Anne Boyer
A good way to help users finding relevant items on document platforms consists in suggesting content in accordance with their preferences. When implementing such a recommender system, the number of potential users and the confidential nature of some data should be taken into account. This paper introduces a new P2P recommender system which models individual preferences and exploits them through a user-centered filtering algorithm. The latter has been designed to deal with problems of scalability, reactivity, and privacy.
Multiple Evidence Combination in Web Site Search Based on Users' Access Histories BIBAFull-Text 405-409
  Chen Ding; Jin Zhou
Despite the success of global search engines, web site search is still problematic in its retrieval accuracy. In this study, we propose to extract terms based on users' access histories to build web page representations, and then use multiple evidence combination to combine these log-based terms with text-based and anchor-based terms. We test different combination approaches and baseline retrieval models. Our experimental results show that the server log, when used in multiple evidence combination, can improve the effectiveness of the web site search, whereas the impact on different models is different.
MyPlace Locator: Flexible Sharing of Location Information BIBAKFull-Text 410-414
  Mark Assad; David J. Carmichael; Judy Kay; Bob Kummerfeld
As location information plays such an important role in pervasive and context-aware computing, location modelling can be cast as a particularly important user modelling problem. Moreover, given the potential sensitivity of personal information about location, it is critical to ensure adequate user control over the use of the location user modelling information. This paper describes MyPlace Locator, a system for modelling people's location, based upon a range of pervasive sensors. A critical feature, the focus of this paper, is the user model control: users can determine the granularity in space and time of the location information released from their model. Users can do this on the level of a single user or a group of users. We describe the interface and report its qualitative evaluation.
Keywords: location modelling; pervasive computing; user control; scrutability
Personalised Mashups: Opportunities and Challenges for User Modelling BIBAKFull-Text 415-419
  Minh Dang Thang; Vania Dimitrova; Karim Djemame
Web 2.0 has emerged as the business ideology and development paradigm for the next generation of web applications. This paper proposes the use of personalisation techniques to enhance the functionality of web mashups, one of the most popular Web 2.0 applications. A prototype of a personalised travel assistant which combines interactive maps with public data pulled from the Internet is presented. An experimental study with the prototype points at opportunities and challenges mashups bring to personalisation research.
Keywords: Web 2.0; mashups; application of user modelling
Personalized Control of Smart Environments BIBAFull-Text 420-424
  Giovanni Cozzolongo; Berardina De Carolis; Sebastiano Pizzutilo
Interaction with smart environments, to be effective, should be easy, natural and should be proactively adapted to users needs. In this paper we propose the use of a butler agent acting as a mediator between environment devices and users. As any good butler, it is able to observe and learn about users preferences but it leaves to its "owner" the last word on decisions. This is possible by employing user and context modeling techniques in order to provide a dynamic adaptation of the interaction with the environment according to the vision of ambient intelligence.
Studying Model Ambiguity in a Language ITS BIBAKFull-Text 425-429
  Brent Martin; Amanda Nicholas
Ambiguity is a well-known problem in student modelling, and in user modelling in general. In this paper we present the results of an experiment in the domain of German adjectives. We trialled a modified student interface that gathers more data during problem solving by requiring the student to perform a related subtask. There is evidence that the students who performed the subtask outperformed the control group on a post-test despite the extra task slowing them down, suggesting the extra effort required by the students to overcome ambiguity was worth the intervention.
Keywords: Student Modelling; Language learning; ITS
Tailoring and the Efficiency of Information Seeking BIBAFull-Text 430-434
  Nathalie Colineau; Cécile Paris
We present an empirical study assessing the impact of tailoring on information seeking tasks. Our aim was to evaluate whether providing tailored information would help people find the information they need more quickly and more accurately. Our results show that tailored documents have an impact on information seeking, at least when the information to be found is spread over a number of sources and needs to be synthesised.
The Effect of Model Granularity on Student Performance Prediction Using Bayesian Networks BIBAFull-Text 435-439
  Zachary A. Pardos; Neil T. Heffernan; Brigham Anderson; Cristina L. Heffernan
A standing question in the field of Intelligent Tutoring Systems and User Modeling in general is what is the appropriate level of model granularity (how many skills to model) and how is that granularity derived? In this paper we will explore models with varying levels of skill generality (1, 5, 39 and 106 skill models) and measure the accuracy of these models by predicting student performance within our tutoring system called ASSISTment as well as their performance on a state standardized test. We employ the use of Bayes nets to model user knowledge and to use for prediction of student responses. Our results show that the finer the granularity of the skill model, the better we can predict student performance for our online data. However, for the standardized test data we received, it was the 39 skill model that performed the best. We view this as support for fine-grained skill models despite the finest grain model not predicting the state test scores the best.
To Share or Not to Share: Supporting the User Decision in Mobile Social Software Applications BIBAKFull-Text 440-444
  Giuseppe Lugano; Pertti Saariluoma
User's privacy concerns represent one of the most serious obstacles to the wide adoption of mobile social software applications. In this paper, we introduce a conceptual model which tackles the problem from the perspective of trade-off between privacy and trust, where the user takes the decision with minimal privacy loss. To support the user decision, we introduce the Mobile Access Control List (Macl), a privacy management mechanism which takes into account the user attitude towards mobile sharing, his communication history and social network relationships.
Keywords: Privacy; Sharing; Trust; Mobile Social Software
Towards a Tag-Based User Model: How Can User Model Benefit from Tags? BIBAFull-Text 445-449
  Francesca Carmagnola; Federica Cena; Omar Cortassa; Cristina Gena; Ilaria Torre
Social tagging is a kind of social annotation by which users label resources, typically web objects, by means of keywords with the goal of sharing, discovering and recovering them. In this paper we investigate the possibility of exploiting the user tagging activity in order to infer knowledge about the user. Up to now the relation between tagging and user modeling seems not to have been investigated in depth. Given the widespread diffusion of web tools for collaborative tagging, it is interesting to understand how user modeling can benefit from this feedback.
Web Customer Modeling for Automated Session Prioritization on High Traffic Sites BIBAKFull-Text 450-454
  Nicolas Poggi; Toni Moreno; Josep Lluis Berral; Ricard Gavaldà; Jordi Torres
In the Web environment, user identification is becoming a major challenge for admission control systems on high traffic sites. When a web server is overloaded there is a significant loss of throughput when we compare finished sessions and the number of responses per second; longer sessions are usually the ones ending in sales but also the most sensitive to load failures. Session-based admission control systems maintain a high QoS for a limited number of sessions, but does not maximize revenue as it treats all non-logged sessions the same. We present a novel method for learning to assign priorities to sessions according to the revenue that will generate. For this, we use traditional machine learning techniques and Markov-chain models. We are able to train a system to estimate the probability of the user's purchasing intentions according to its early navigation clicks and other static information. The predictions can be used by admission control systems to prioritize sessions or deny them if no resources are available, thus improving sales throughput per unit of time for a given infrastructure. We test our approach on access logs obtained from a high-traffic online travel agency, with promising results.
Keywords: Web prediction; navigation patterns; machine learning; data mining; admission control; resource management; autonomic computing; e-commerce
What's in a Step? Toward General, Abstract Representations of Tutoring System Log Data BIBAKFull-Text 455-459
  Kurt VanLehn; Kenneth R. Koedinger; Alida Skogsholm; Adaeze Nwaigwe
The Pittsburgh Science of Learning Center (PSLC) is developing a data storage and analysis facility, called DataShop. It currently handles log data from 6 full-year tutoring systems and dozens of smaller, experimental tutoring systems. DataShop requires a representation of log data that supports a variety of tutoring systems, atheoretical analyses and theoretical analyses. The theory-based analyses are strongly related to student modeling, so the lessons learned in developing the DataShop's representation may apply to student modeling in general. This report discusses the representation originally used by the DataShop, the problems encountered, and how the key concept of "step" evolved to meet these challenges.
Keywords: Student modeling; educational data mining; tutoring systems

Doctoral Consortium Papers

Encouraging Contributions to Online Communities with Personalization and Incentives BIBAKFull-Text 460-464
  F. Maxwell Harper
Increasingly, online systems depend on user contributions such as posts, ratings, tags, and comments. Many of these systems wish to encourage broader participation or the contribution of higher quality content. In this doctoral consortium paper, I present past work and propose future work on understanding user motivations to contribute online and on the use of personalization technology and incentives to shape participation.
Keywords: Incentives; personalization; online communities
Semantic-Enhanced Personalised Support for Knowledge Sharing in Virtual Communities BIBAFull-Text 465-469
  Styliani Kleanthous
Virtual communities are currently one of the fastest growing applications on the web. In this research, we argue that personalised support should be tailored to the needs of the community as a whole, as opposed to adapting only to individuals. Based on 4 processes identified as important, we propose a computational framework that includes the extraction of a comprehensive community model and the deployment of that model to provide support adapted to the effective functioning of a community.
Explaining Recommendations BIBAFull-Text 470-474
  Nava Tintarev
This thesis investigates the properties of a good explanation in a movie recommender system. Beginning with a summarized literature review, we suggest seven criteria for evaluation of explanations in recommender systems. This is followed by an attempt to define the properties of a useful explanation, using a movie review corpus and focus groups. We conclude with planned experiments and evaluation.
Designing Persuasive Health Behaviour Change Dialogs BIBAFull-Text 475-479
  Hien Nguyen
Using theories of behaviour change, argumentation theory, and findings in social psychology, our research explores new methods to raise the persuasiveness of adaptive dialog-based systems using tailored arguments and onscreen characters to enhance the system's credibility and trustworthiness. Initial results revealed the existence of individual preferences for arguments, types of communication, and appearance of onscreen characters. In the future, we will explore methods to learn these preferences through interactions with the user, and to utilize them to maximize the persuasion effect of the system. The final outcome of the research will be a persuasion model that is capable of modelling the user's cognitive and affective state and generating tailored arguments to move the user in the desired direction.
User-Centered Design for Personalized Access to Cultural Heritage BIBAKFull-Text 480-484
  Yiwen Wang
The volume of digital cultural heritage is huge and rapidly growing. The overload of art information has created the need to help people find out what they like in the enormous museum collections and provide them with the most convenient access point. In this paper, we present a research plan to address these issues. Our approach involves: (1) use of ontologies as shared vocabularies and thesauri to model the domain of art; (2) an interactive ontology-based elicitation of user interests and preferences in art to be stored as an extended overlay user model; (3) RDF/OWL reasoning strategies for predicting users' interests and generating recommendations; and (4) The Rijksmuseum Amsterdam use case for a personalized museum tour combining both the virtual Web space and the physical museum space to enhance the users' experience. We follow a user-centered design for collecting requirements, testing out design choices and evaluating stages of our prototypes.
Keywords: CHIP (Cultural Heritage Information Presentation); user-centered design; user modeling; personalization; Semantic Web; RDF; recommendations