scispace - formally typeset
Search or ask a question

Showing papers on "User modeling published in 1999"


Patent
Dan Newell1, Kenneth H. Abbott1
15 Dec 1999
TL;DR: In this paper, the authors present a system for controlling presentation of information to a user based on the user's current condition, i.e., cognitive load, desired level of privacy, and desired scope of audience for output information.
Abstract: A system for controlling presentation of information to a user based on the user's current condition. In particular, the system monitors the user and the user's environment, and creates and maintains an updated model of the current condition of the user. The user condition can include a variety of condition variables, including abstract concepts such as the user's current cognitive load, desired level of privacy for output information, and desired scope of audience for output information. Upon receiving output information to be presented to the user (e.g., from an application program), the system determines an appropriate output device and an appropriate format with which to present the information to the user, and then presents the output information. The system can also receive description information about the output information that describes relevant factors for determining when and how to present the output information (e.g., the importance and urgency of the output information, the consequences of the user not receiving or ignoring the output information, etc.). Some versions of the system execute on a wearable computer having a variety of available output display devices.

566 citations


Journal ArticleDOI
TL;DR: This article presents some situations in which individual user models are important and some techniques that make the construction and use of such models possible and the performance of one system that uses some of these techniques is discussed.
Abstract: It has long been recognized that in order to build a good system in which a person and a machine cooperate to perform a task it is important to take into account some significant characteristics of people. These characteristics are used to build some kind of a “user model”. Traditionally, the model that is built is a model of a canonical (or typical) user. But often individual users vary so much that a model of a canonical user is insufficient. Instead, models of individual users are necessary. This article presents some examples of situations in which individual user models are important. It also presents some techniques that make the construction and use of such models possible. These techniques all reflect a desire to place most of the burden of constructing the models on the system, rather than on the user. This leads to the development of models that are collections of good guesses about the user. Thus some kind of probabilistic reasoning is necessary. And as the models are being used to guide the underlying system, they must also be monitored and updated as suggested by the interactions between the user and the system. The performance of one system that uses some of these techniques is discussed.

497 citations


Proceedings ArticleDOI
01 Feb 1999
TL;DR: A reference model for adaptive hypermedia applications, called AHAM, is described, which encompasses most features supported by adaptive systems that exist today or that are being developed (and have been published about).
Abstract: Hypermedia applications offer users the impression that there are many meaningful ways to navigate through a large body of information nodes. This rich link structure not only creates orientation problems, it may also be a source of comprehension problems when users follow paths through the information which the author did not foresee. Adaptive techniques have been used by a number of researchers [1, 2, 4, 5, 6, 7, 8, 9, 10, 17, 19, 20, 22] in an attempt to offer guidance through and orientation support for rich link structures. The majority of these adaptive hypermedia systems (AHS) have been used in educational applications. The terminology used in this paper also has an educational “flavor”. However, there are some adaptive on-line information systems (or “kiosk”systems), adaptive information retrieval systems, and other adaptive hypermedia applications. In this paper we describe a reference model for adaptive hypermedia applications, called AHAM, which encompasses most features supported by adaptive systems that exist today or that are being developed (and have been published about). Our description of AHS is based on the Dexter model [15, 16], a widely used reference model for hypermedia. The description is kept somewhat informal in order to be able to explain AHAM rather than formally specify it. AHAM augments Dexter with features for doing adaptation based on a user model which persists beyond the duration of a session. Key aspects in AHAM are: Paul De Bra is also affiliated with the University of Antwerp, Belgium, and with the “Centrum voor Wiskunde en Informatica” (CWI) in Amsterdam. yGeert-Jan Houben is also affiliated with the University of Antwerp, Belgium, and with Origin in Eindhoven. The adaptation is based on a domain model, a user model and a teaching model which consists of pedagogical rules. We give a formal definition of each of these (sub)models (but only describe the pedagogical rules informally throughexamples). We distinguish the notions of concept, page and fragment. In some AHS these notions are confused. We provide a formalism which lets authors write pedagogical rules (about concepts) in such a way that they can be applied automatically. We illustrate various aspects of AHAM by means of some features of some well-known AHS [6, 10].

494 citations


Patent
09 Jun 1999
TL;DR: In this article, a location-based discrimination mechanism is proposed to distinguish local users, intranet users, and dial-up users from one another, based on information including the location and user's credentials.
Abstract: An improved computer network security system and method wherein access to network resources is based on information that includes the location of the connecting user. In general, the less trusted the location of the user, the more the access rights assigned to the user are restricted. A discrimination mechanism and process determines the location of a user with respect to categories of a security policy, such as to distinguish local users, intranet users and dial-up users from one another. Based on information including the location and the user's credentials, an access token is set up that may restrict the user's normal access in accordance with the security policy, such as to not restrict a user's processes beyond the user-based security information in the user's normal access token, while further restricting the same user's access to resources when connecting via a dial-up connection. Restricted tokens are preferably used to implement the location-based discrimination by restricting the security context of users connecting from less trusted locations.

436 citations


Book ChapterDOI
01 Jun 1999
TL;DR: An intelligent agent designed to compile a daily news program for individual users, which motivates the use of a multi-strategy machine learning approach that allows for the induction of user models that consist of separate models for long-term and short-term interests.
Abstract: We present an intelligent agent designed to compile a daily news program for individual users. Based on feedback from the user, the system automatically adapts to the user’s preferences and interests. In this paper we focus on the system’s user modeling component. First, we motivate the use of a multi-strategy machine learning approach that allows for the induction of user models that consist of separate models for long-term and short-term interests. Second, we investigate the utility of explicitly modeling information that the system has already presented to the user. This allows us to address an important issue that has thus far received virtually no attention in the Information Retrieval community: the fact that a user’s information need changes as a direct result of interaction with information. We evaluate the proposed algorithms on user data collected with a prototype of our system, and assess the individual performance contributions of both model components.

427 citations


Journal ArticleDOI
TL;DR: An approach that transforms temporal sequences of discrete, unordered observations into a metric space via a similarity measure that encodes intra-attribute dependencies and demonstrates that it can accurately differentiate the profiled user from alternative users when the available features encode sufficient information.
Abstract: The anomaly-detection problem can be formulated as one of learning to characterize the behaviors of an individual, system, or network in terms of temporal sequences of discrete data. We present an approach on the basis of instance-based learning (IBL) techniques. To cast the anomaly-detection task in an IBL framework, we employ an approach that transforms temporal sequences of discrete, unordered observations into a metric space via a similarity measure that encodes intra-attribute dependencies. Classification boundaries are selected from an a posteriori characterization of valid user behaviors, coupled with a domain heuristic. An empirical evaluation of the approach on user command data demonstrates that we can accurately differentiate the profiled user from alternative users when the available features encode sufficient information. Furthermore, we demonstrate that the system detects anomalous conditions quickly — an important quality for reducing potential damage by a malicious user. We present several techniques for reducing data storage requirements of the user profile, including instance-selection methods and clustering. As empirical evaluation shows that a new greedy clustering algorithm reduces the size of the user model by 70%, with only a small loss in accuracy.

395 citations


Journal ArticleDOI
TL;DR: A structured, iterative methodology for user-centered design and evaluation of VE user interaction is presented and why this approach provides a cost-effective strategy for assessing and iteratively improving user interaction in VEs is evaluated.
Abstract: We present a structured, iterative methodology for user-centered design and evaluation of VE user interaction. We recommend performing (1) user task analysis followed by (2) expert guidelines-based evaluation, (3) formative user-centered evaluation, and finally (4) comparative evaluation. In this article we first give the motivation and background for our methodology, then we describe each technique in some detail. We applied these techniques to a real-world battlefield visualization VE. Finally, we evaluate why this approach provides a cost-effective strategy for assessing and iteratively improving user interaction in VEs.

341 citations


Proceedings Article
01 Jan 1999
TL;DR: This paper presents a generic framework inspired by the model-based approach, for supporting the development of plastic user interfaces, and aims to guarantee usability continuity under variations in physical constraints while minimizing development and maintenance costs.
Abstract: This paper introduces the notion of plasticity, a new property of interactive systems that denotes a particular type of user interface adaptation Plasticity is the capacity of a user interface to withstand variations of both the system physical characteristics and the environment while preserving usability Typically, a ‘plastic’ electronic agenda would run both on a workstation and on a hand-held computer without requiring a complete system redesign and re-implementation We present a generic framework inspired by the model-based approach, for supporting the development of plastic user interfaces Within this framework, a plastic user interface is specified once and serves multiple sources of physical variations The goal is to guarantee usability continuity under variations in physical constraints while minimizing development and maintenance costs This framework is illustrated with two simple case studies Preliminary results and the state of the art in HCI open a new research agenda for the design and development of plastic user interfaces

338 citations


Patent
28 Jun 1999
TL;DR: In this article, an automated recommendation system keeps track of the needs and preferences of the user through a user preference vector, which represents the user's preference for a specific item attribute.
Abstract: An automated recommendation system keeps track of the needs and preferences of the user through a user preference vector. Each field of the user preference vector represents the user's preference for a specific item attribute. Item attributes are defined by a systems programmer. The systems programmer also creates product vectors of items in the recommendation database. A user preference vector is compared against a product vector to determine if the product is suitable for recommendation. A recommended item may be purchased by the user by submitting a purchase request over a network connection. The user preference vector is constantly refined through feedback from the user about the recommended items.

329 citations


Journal ArticleDOI
TL;DR: A framework for adaptive hypermedia systems (AHS) is introduced and some popular methods and techniques for adaptation are described, to illustrate the potential benefits of using adaptation in hypermedia applications.
Abstract: The navigational freedom in conventional hypermedia applications leads to comprehension and orientation problems (Nielsen 1990). Adaptive hypermedia attempts to overcome these problems by adapting the presentation of information and the overall link structure, based on a user model. This paper introduces a framework for adaptive hypermedia systems (AHS). It briefly describes some popular methods and techniques for adaptation. Examples and evaluations of existing AHS are used to illustrate the potential benefits of using adaptation in hypermedia applications.

318 citations


Book ChapterDOI
01 Jun 1999
TL;DR: A large corpus of Web search queries extracted from server logs recorded by a popular Internet search service is derived and Bayesian networks that predict search behavior are constructed, with a focus on the progression of queries over time.
Abstract: We discuss the construction of probabilistic models centering on temporal patterns of query refinement. Our analyses are derived from a large corpus of Web search queries extracted from server logs recorded by a popular Internet search service. We frame the modeling task in terms of pursuing an understanding of probabilistic relationships among temporal patterns of activity, informational goals, and classes of query refinement. We construct Bayesian networks that predict search behavior, with a focus on the progression of queries over time. We review a methodology for abstracting and tagging user queries. After presenting key statistics on query length, query frequency, and informational goals, we describe user models that capture the dynamics of query refinement.

Proceedings ArticleDOI
06 Nov 1999
TL;DR: In this article, the starER model combines the star structure, which is dominant in data warehouses, with the semantically rich constructs of the ER model; special types of relationships have been further added to support hierarchies.
Abstract: Modeling data warehouses is a complex task focusing, very often, into internal structures and implementation issues. In this paper we argue that, in order to accurately reflect the users requirements into an error-free, understandable, and easily extendable data warehouse schema, special attention should be paid at the conceptual modeling phase. Based on a real mortgage business warehouse environment, we present a set of user modeling requirements and we discuss the involved concepts. Understanding the semantics of these concepts, allow us to build a conceptual model—namely, the starER model—for their efficient handling. More specifically, the starER model combines the star structure, which is dominant in data warehouses, with the semantically rich constructs of the ER model; special types of relationships have been further added to support hierarchies. We present an evaluation of the starER model as well as a comparison of the proposed model with other existing models, pointing out differences and similarities. Examples from a mortgage data warehouse environment, in which starER is tested, reveal the ease of understanding of the model, as well as the efficiency in representing complex information at the semantic level.

Patent
21 Jun 1999
TL;DR: In this article, a testing tool automatically records a series of user steps taken during a user session with a transactional server and generates a test for testing the functionality of server through a user interface.
Abstract: A testing tool automatically records a series of user steps taken during a user session with a transactional server and generates a test for testing the functionality of server. Through a user interface of the testing tool, the user can define verification steps to automatically test for expected server responses during test execution. The testing tool displays the test to the user as a tree having nodes (displayed as icons) which represent steps of the test. Via the user interface, the user can modify node properties and perform other types of tree edit operations to edit the test, without the need to know a scripting or other programming language. When the user selects a node that corresponds to a particular field or other object of the server screen, the testing tool automatically displays the screen with the object highlighted. The testing tool also allows the test author to use a spreadsheet to conveniently specify data sets for running multiple iterations of a test; thus, the user can record a single transaction and then automatically test the transaction with other data sets.

Proceedings ArticleDOI
01 Apr 1999
TL;DR: A personal news agent is presented that is designed to become part of an intelligent, IP-enabled radio, which uses synthesized speech to read news stories to a user based on voice feedback from the user, which automatically adapts to the user’s preferences and interests.
Abstract: Most work on intelligent information agents has thus far focused on systems that are accessible through the World Wide Web. As demanding schedules prohibit people from continuous access to their computers, there is a clear demand for information systems that do not require workstation access or graphical user interfaces. We present a personal news agent that is designed to become part of an intelligent, IP-enabled radio, which uses synthesized speech to read news stories to a user. Based on voice feedback from the user, the system automatically adapts to the user’s preferences and interests. In addition to time-coded feedback, we explore two components of the system that facilitate the automated induction of accurate interest profiles. First, we motivate the use of a multistrategy machine learning approach that allows for the induction of user models that consist of separate models for long-term and short-term interests. Second, we investigate the use of "concept feedback", a novel form of user feedback that is based on our agent's capability to construct explanations for the reasons that have led to a specific classification. Users can then critique these explanations which, from a machine learning perspective, allows for more direct changes to an induced concept than through the inclusion of additional training examples. We evaluate the proposed algorithms on user data collected with a prototype of our system, and assess the performance contributions of the system's individual components.

Patent
10 Sep 1999
TL;DR: In this paper, a system and methods for creating and delivering interactive multimedia content that can dynamically adapt to at least one user is presented. But it is not shown to the entire audience.
Abstract: System and methods for creating and delivering interactive multimedia content that can dynamically adapt to at least one user. At least one user model is created for at least one user. The at least one user model represents interests and trends of the at least one user. A multimedia story is developed based on the at least one user model. A customized presentation of the multimedia story is generated. The at least one multimedia story allows for multiple presentations of the multimedia story. The customized presentation is displayed to the at least one user. The customized presentation is modified based on input from the at least one user.

Patent
Chanda Dharap1
10 Jun 1999
TL;DR: In this article, a context is created based on a profile of the user, the profile being at least partly formed in advance and candidate data is selected from the data base under control of the context and the user is enabled to interact with the candidates.
Abstract: A user is enabled to navigate through an electronic data base in a personalized manner. A context is created based on a profile of the user, the profile being at least partly formed in advance. Candidate data is selected from the data base under control of the context and the user is enabled to interact with the candidates. The profile is based on topical information supplied by the user in advance and a history of previous accesses from the user to the data base.

Book ChapterDOI
Pat Langley1
01 Jun 1999
TL;DR: The notion of adaptive user interfaces, interactive systems that invoke machine learning to improve their interaction with humans, is examined and three ongoing research efforts that extend this framework in new directions are described.
Abstract: In this paper we examine the notion of adaptive user interfaces, interactive systems that invoke machine learning to improve their interaction with humans. We review some previous work in this emerging area, ranging from software that filters information to systems that support more complex tasks like scheduling. After this, we describe three ongoing research efforts that extend this framework in new directions. Finally, we review previous work that has addressed similar issues and consider some challenges that are presented by the design of adaptive user interfaces.

Patent
15 Apr 1999
TL;DR: In this paper, a system for allowing a computer network site to recognize an anonymous user without revealing the identity of the user is proposed, which involves generating a user alias based on the user identity and the computer network sites such that it is computationally difficult to determine the user's identity from the alias alone.
Abstract: An system for allowing a computer network site to recognize an anonymous user without revealing the identity of the user. The system involves generating a user alias based on the user's identity and the computer network site such that it is computationally difficult to determine the user's identity from the alias alone. The system further involves informing the computer network site of the alias upon access of the site by the user. The computer network site may then block access to the site's contents whenever it receives an alias associated with a disruptive user.

Patent
08 Jan 1999
TL;DR: In this paper, a system for automatically generating user interest profiles and delivering information to users learns a user's interests by monitoring the user's outbound communication streams, i.e., the information that the user produces either by typing or speaking.
Abstract: A system for automatically generating user interest profiles and delivering information to users learns a user's interests by monitoring the user's outbound communication streams, i.e., the information that the user produces either by typing (e.g., while a user is composing an e-mail message or editing a word processor document) or by speaking (e.g., while a user is engaged in a phone conversation or listening to a lecture). The system uses the monitored text to build (and possibly update) a user interest profile. The profile is constructed from current text generated by the user, so that the retrieved information reflects present user interests. In addition, the profile may also retain past user interests, so that the profile reflects a combination of past and present user interests. The system then automatically queries diverse databases for information relevant to the interest profile. The databases may include internet web pages, files stored on the user's local network, and other local or remote data repositories. The queries may use a combination of internet search engines, the specific selection of which may depend upon the nature and/or content of the queries. The information retrieved in response to the queries is then presented to the user. The retrieved information may contain, for example, answers to questions that the user might ask and/or data related to the user's current and continuing interests. Because a user's current speech or typed text is highly correlated with the user's current interests, the retrieved information will be relevant to the user's actual interests. The communication stream monitoring, interest profile building, data base querying, and presentation of retrieved information are all performed automatically, in real time, and in the background of current user activities.

Proceedings ArticleDOI
01 Nov 1999
TL;DR: A new scheme to learn dynamic user's interests in an automated information filtering and gathering system running on the Internet that adapts quickly to significant changes in user interest, and is also able to learn exceptions to interest categories.
Abstract: In this paper, we describe a new scheme to learn dynamic user's interests in an automated information filtering and gathering system running on the Internet. Our scheme is aimed to handle multiple domains of long-term and short-term user's interests simultaneously, which is learned through positive and negative user's relevance feedback. We developed a 3-descriptor approach to represent the user's interest categories. Using a learning algorithm derived for this representation, our scheme adapts quickly to significant changes in user interest, and is also able to learn exceptions to interest categories.

Patent
16 Aug 1999
TL;DR: In this article, a client computer program provides instructions for a processor to locate and retrieve rich media and HTML files for running in a 3D graphical user interface (GUI) on a display, and then the processor generates a metaphorical user object for navigating and interacting in the three dimensions within the environment via navigational and interactional inputs, respectively, from a user.
Abstract: A client computer program provides instructions for a processor to locate and retrieve rich media and HTML files for running in a three-dimensional graphical user interface. The program provides instructions for the processor to generate a three-dimensional (3D) graphical user interface on a display. Then, the processor generates a metaphorical user object for navigating and interacting in the three dimensions within the environment via navigational and interactional inputs, respectively, from a user. The processor also integrates an HTML object into the environment, and interacts the user object with the HTML object in response to an interactional input from the user. Preferably, the processor generates one or more additional user objects for additional users. Each of the additional user objects can also interact with the HTML object, the first user object and/or each other within the environment in response to interactional inputs from the additional users. In addition, the first user object may interact with any or all of the additional user objects in response to interactional inputs from the first user. The interactional inputs may be proxy expressions from one user to another.

Patent
26 Jan 1999
TL;DR: In this paper, a wizard-style interface is used to establish a study plan for a user based on a session frequency, along with the date the user wishes to finish the instructional product or the amount of time that the user has available each session.
Abstract: A method and system for setting up and using a personal learning schedule and path through an instructional product. Using a wizard-style interface, a study plan is established for a user based on a session frequency, along with the date the user wishes to finish the instructional product or the amount of time that the user has available each session. Once the user has set up a study plan, each time that user logs in, the user is taken to that day's list of assigned activities on a personalized home page. From there, the user is able to check activity status for past and future lessons, and quickly navigate to a lesson. Users can set up their own personalized learning schedule based on their own study objectives. The method and system further allow the user to determine a skill level required to complete the activities, and further monitors the user's progress in successfully completing those activities.

Book ChapterDOI
22 Sep 1999
TL;DR: The issue of presenting a general user profile model for digital libraries users is addressed and the general user profiles representing users' information needs are customised fordigital libraries users.
Abstract: The ultimate goal of an information provider is to satisfy the user information needs That is, to provide the user with the right information, at the right time, through the right means A prerequisite for developing personalised services is to rely on user profiles representing users' information needs In this paper we will first address the issue of presenting a general user profile model Then, the general user profile model will be customised for digital libraries users

Patent
07 Dec 1999
TL;DR: In this paper, a user interface architecture consisting of user content and user interface is composed of documents with links, and a non-linear navigation history is maintained such that a user can navigate along a first path, back up using a previous link, navigate along another path, and re-navigate along the first path again using a next link.
Abstract: A user interface architecture wherein user content and user interface are composed of documents with links. Links have properties that indicate how links should be displayed. Link display may depend upon characteristics of the linked-to frame. Commands are implemented as links that link to command code. The display format of a document page depends upon the path a user navigated in order to get to the displayed page. User content documents and user interface documents are both displayed in a single viewing frame. A non-linear navigation history is maintained such that a user can navigate along a first path, back up using a previous link, navigate along a second path, back up along the second path using the previous link, and re-navigate along the first path again using a next link. Every document page to which a user navigates is saved in the user's navigation history. Users can query their navigation histories and view their navigation histories in various ways. A flexible selection model is supported allowing users to select the object of a command either before or after the command itself is selected. A Guide Book includes command descriptions and active instances of the described commands, implemented as links. Users can access, annotate, and customize the user interface in the same manner in which users access, annotate, and modify user content documents.

Journal ArticleDOI
TL;DR: A method to structure use-case models with goals is developed based on industry trends and research and uses a simple meeting planner system to illustrate the benefits of this new approach.
Abstract: The purpose of requirements engineering is to elicit and evaluate necessary and valuable user needs. Current use-case approaches to requirements acquisition inadequately support use-case formalization and nonfunctional requirements. Based on industry trends and research, the authors have developed a method to structure use-case models with goals. They use a simple meeting planner system to illustrate the benefits of this new approach.

Book ChapterDOI
01 Jun 1999
TL;DR: This paper synthesizes results of previous experimental studies of the ways in which a speaker’s cognitive load is reflected in features of speech and presents a conceptualization of these relationships in terms of Bayesian networks.
Abstract: Users of computing devices are increasingly likely to be subject to situationally determined distractions that produce exceptionally high cognitive load. The question arises of how a system can automatically interpret symptoms of such cognitive load in the user’s behavior. This paper examines this question with respect to systems that process speech input. First, we synthesize results of previous experimental studies of the ways in which a speaker’s cognitive load is reflected in features of speech. Then we present a conceptualization of these relationships in terms of Bayesian networks. For two examples of such symptoms—sentence fragments and articulation rate—we present results concerning the distribution of the symptoms in realistic assistance dialogs. Finally, using artificial data generated in accordance with the preceding analyses, we examine the ability of a Bayesian network to assess a user’s cognitive load on the basis of limited observations involving these two symptoms.

Patent
13 Oct 1999
TL;DR: In this paper, the authors propose a front end to a database search engine or engines that process a user query to generate a new search request that will more effectively retrieve information from the database that is relevant to the query of the user.
Abstract: The invention provides, inter alia, front ends to a database search engine or engines, that process a user query to generate a new search request that will more effectively retrieve information from the database that is relevant to the query of the user. To this end, in one embodiment the systems can be realized as computer programs present to a user interface to a user and which prompt the user to enter one or more key phrases that are representative of a user search request or user query. The user interface can collect the key phrases provided by the user and can analyze these key phrases to identify at least one meaning that can be associated with this user query. The systems can then process the user query and the identified meaning to generate an expanded search request that can be represented as a boolean search strategy. This boolean search strategy can then be processed to create one or more expanded user queries that can be presented to a search engine to collect from a search engine information that is relevant to the interest of the user.

Book ChapterDOI
01 Jun 1999
TL;DR: The user modeling and personalization techniques adopted in SETA, a shell supporting the construction of adaptive Web stores which customize the interactions with users, suggesting the items best fitting their needs, and adapting the description of the store catalog to their preferences and expertise are described.
Abstract: We describe the user modeling and personalization techniques adopted in SETA, a shell supporting the construction of adaptive Web stores which customize the interactions with users, suggesting the items best fitting their needs, and adapting the description of the store catalog to their preferences and expertise. SETA uses stereotypical information to handle the user models and applies personalization rules to dynamically generate the hypertextual pages presenting products: the system adapts the graphical aspect, length and terminology used in the descriptions to the user’s receptivity, expertise and interests. Moreover, it maintains a profile associated to each person the goods are selected for, to provide multiple criteria for the selection of items, tailored to the beneficiaries’ preferences.

Book ChapterDOI
01 Jun 1999
TL;DR: The encouraging results presented in this paper, suggest that established machine learning methods can be particularly useful for the acquisition of stereotypes and communities.
Abstract: In this paper we examine the acquisition of user stereotypes and communities automatically from users’ data. Stereotypes are built using supervised learning (C4.5) on personal data extracted from a set of questionnaires answered by the users of a news filtering system. Particular emphasis is given to the characteristic features of the task of learning stereotypes and, in this context, the new notion of community stereotype is introduced. On the other hand, the communities are built using unsupervised learning (COBWEB) on data containing users’ interests on the news categories covered by the news filtering system. Our main concern is whether meaningful communities can be constructed and for this purpose we specify a metric to decide on the representative news categories for each community. The encouraging results presented in this paper, suggest that established machine learning methods can be particularly useful for the acquisition of stereotypes and communities.

Book ChapterDOI
01 Aug 1999
TL;DR: The paper motivates the AREA framework and discusses the system in terms of its formal modeling capabilities and its operational aspects, followed by practical implementation considerations addressing the role of AREA as a groupware infrastructure.
Abstract: This paper presents AREA, an integrated synchronous and asynchronous notification service for awareness information. AREA uses a semantic model of the client applications to support cross-application awareness. The service is based on the dichotomy of interest and privacy. Notifications of user activities are a function of relevance in the current work situation and the privacy requirements of the involved users. The paper motivates the AREA framework and discusses the system in terms of its formal modeling capabilities and its operational aspects. This is followed by practical implementation considerations addressing the role of AREA as a groupware infrastructure. Finally, a prototype groupware application is presented, which uses the AREA service to support user awareness.