This book constitutes the refereed proceedings of the Third International Workshop on Cooperative Information Systems, CIA'99, held in Uppsala, Sweden in July/August 1999.
The 16 revised full papers presented were carefully reviewed and selected from a total of 46 submissions. Also included are ten invited contributions by leading experts. The volume is divided in sections on information discovery and management on the Internet; information agents on the Internet-prototypes systems and applications; communication and collaboration, mobile information agents; rational information agents for electronic business; service mediation and negotiation; and adaptive personal assistance.

Past, Present, and Future of Knowledge Acquisition This book contains the proceedings of the 11th European Workshop on Kno- edge Acquisition, Modeling, and Management (EKAW '99), held at Dagstuhl Castle (Germany) in May of 1999. This continuity and the high number of s- missions re?ect the mature status of the knowledge acquisition community. Knowledge Acquisition started as an attempt to solve the main bottleneck in developing expert systems (now called knowledge-based systems): Acquiring knowledgefromahumanexpert. Variousmethodsandtoolshavebeendeveloped to improve this process. These approaches signi?cantly reduced the cost of - veloping knowledge-based systems. However, these systems often only partially ful?lled the taskthey weredevelopedfor andmaintenanceremainedanunsolved problem. This required a paradigm shift that views the development process of knowledge-based systems as a modeling activity. Instead of simply transf- ring human knowledge into machine-readable code, building a knowledge-based system is now viewed as a modeling activity. A so-called knowledge model is constructed in interaction with users and experts. This model need not nec- sarily re?ect the already available human expertise. Instead it should provide a knowledgelevelcharacterizationof the knowledgethat is requiredby the system to solve the application task. Economy and quality in system development and maintainability are achieved by reusable problem-solving methods and onto- gies. The former describe the reasoning process of the knowledge-based system (i. e. , the algorithms it uses) and the latter describe the knowledge structures it uses (i. e. , the data structures). Both abstract from speci?c application and domain speci?c circumstances to enable knowledge reuse.

Autonomousagents andmultiagentsystems are computationalsystems inwhich several (semi-)autonomous agents interact each other or work together to p- form some set of tasks or satisfy some set of goals. These systems may involve computationalagents that are homogeneous or heterogeneous, they may involve activities on the part of agents having common goals or distinct goals, and they may involve participation on the part of humans and intelligent agents. The increase of agent research activities can be observed in the series of ICMAS (- ternational Conference of Multi-Agent Systems). In ICMAS'95,209participants were there. At ICMAS'96, the number increased to282, and atICMAS'98, there were 552participants. Thoughwealready haveseveral workshopsinAsia Paci c countries, such as MACC (Multiagent Systems and Cooperative Computation) in Japan from 1991 and Australian Workshop on Distributed Artic ial Int- ligence from 1995, there is not enough interaction so far among the countries compared to Europe and Americas. PRIMA'98 is the First Pacic Rim International Workshop related to - tonomous agents and multiagent systems. The aim of this workshop is to - courage activities in this e ld, and to bring together Pacic Rim researchers withagents and multiagentissues. The objective of this workshopis to assemble theoreticians and practitioners concerned with developing autonomous agents and multiagent systems. Unlikeusual conferences, however, since this is the rst agent-related international workshop in Asia Paci c countries, participation is byinvitationonlyand is limitedto professionals who have made signi cantc- tributionsto thetopicsoftheworkshop. Thecontributionsincludetechnical p- sentations, progress reports and so on.

Following on from a three-year knowledge management project, seven organisations formed aco-operative group for knowledge management. This group meets through the Knowledge Management Implementers Forum (KMIF). Each of the organisations participating in this work are, by implication, interested in the development of KM. The aims of the forum are t9 exchange ideas and share experience in the areaofknowledge management. The organisations involved are: ~ British Aerospace (Samlesbury) ~ ICI ~ ICL ~ North WestWater ~ IDS Cad-Graphics ~ Liverpool John Moores University ~ NWAIAG (Blackburn College) 1.1 The Organisations Involved Each ofthe organisations has specific reasons for being involved in this project and in KM. The British Aerospace Samlesbury site is a large manufacturing site employing ground breaking technology for Europe's front line military aircraft. The factory works with a well-managed supply chain and works closely with other British Aerospace sites in the manufacture of aircraft components. It has set up a partnership with another Aerospace Company based on exchange of knowledge and therefore needs to value that knowledge. ICI is one of the UK's leading chemical companies and plays on an international stage. Changes in international supply and demand require ICI to respond quickly to market pressures. This means that the company needs to use its knowledge assets in a well managed way and put systems in place that increase the flexibility and ensure the security ofthese important assets.

R. MILNE Intelligent Applications Ltd The papers in this volume are the Application Papers presented at ES98, the Eighteenth International Conference of the British Computer Society's Specialist Group on Expert Systems. This year has been yet another "applications" success for the conference with this volume containing seventeen papers describing either deployed applications or emerging applications. All these documented case studies provide clear evidence of the success of AI technology in solving real business problems. Six of these papers were nominated for the Best Application Award during the review process. These nominations were then reviewed by the members of the Programme Committee to select the winning paper. The papers in the volume were subject to refereeing by at least two referees. All papers which were controversial for some reason were discussed in depth by the Application Programme Committee. Ten referees from the industrial and commercial sector and nine referees from the academic sector assisted me in reviewing the papers. The review form asked the referee to score the papers according to a number of dimensions, to rate it overall, and to offer critical comments to me, and to the authors. It also asks the referee to score their expertise in the area of each paper they review. Only reviews from 'expert' referees are used.

R.G.MILES XHP Consulting Ltd, Gloucester. This book is one of two volumes containing papers for presentation at the British Computer Society Expert Systems 98 conference. This is the annual conference of th the BCS Specialist Group on Expert Systems and is in its 18 year. During its lifetime it has established itself as the premier Expert Systems conference in the UK. The conference is attracting an increasing number of papers world-wide and this year in excess of 70% were from research groups outside the UK. This volume includes all papers accepted for the Technical Stream of Expert Systems 98 and presented at the conference in December 1998. The papers within this stream present innovative, new research work. The companion volume, Applications and Innovations in Expert Systems VI, includes all papers accepted for the application stream of the conference. This stream has become the premier European conference on applications of Expert Systems. The papers accepted for presentation within the Technical Stream cover a broad range of research within Expert Systems and fit into four broad categories: ontological frameworks, knowledge base development, classifiers and neuro-fuzzy systems. The award for best Technical paper has been made to David McSherry, from the University of Ulster, for his paper entitled "Strategic Induction of Decision Trees".

The biennial International Conference on Case-Based Reasoning (ICCBR) - ries, which began in Sesimbra, Portugal, in 1995, was intended to provide an international forum for the best fundamental and applied research in case-based reasoning (CBR). It was hoped that such a forum would encourage the g- wth and rigor of the eld and overcome the previous tendency toward isolated national CBR communities. The foresight of the original ICCBR organizers has been rewarded by the growth of a vigorous and cosmopolitan CBR community. CBR is now widely recognized as a powerful and important computational technique for a wide range of practical applications. By promoting an exchange of ideas among CBR researchers from across the globe, the ICCBR series has facilitated the broader acceptance and use of CBR. ICCBR-99 has continued this tradition by attracting high-quality research and applications papers from around the world. Researchers from 21 countries submitted 80 papers to ICCBR-99. From these submissions, 17 papers were selected for long oral presentation, 7 were accepted for short oral presentation, and 19 papers were accepted as posters. This volume sets forth these 43 papers, which contain both mature work and innovative new ideas.

ThisvolumecontainsaselectionofpaperspresentedattheInternationalConf- ence on Analytic Tableaux and Related Methods (TABLEAUX'99) held on June 7-11, 1999 at the Inn at Saratoga, Saratoga Springs, NY, USA. This conference was the continuation of international meetings on Theorem Proving with A- lytic Tableaux and Related Methods held in Lautenbach near Karlsruhe (1992), Marseille (1993), Abingdon near Oxford (1994), St. Goar near Koblenz (1995), Terrasini near Palermo (1996), Pont-' a-Mousson near Nancy (1997), and Oist- wijk near Tilburg (1998). TABLEAUX'99 marks the ?rst time the conference has been held in North America. Tableau and related methods have been found to be convenient and e?ective for automating deduction in various non-standard logics as well as in classical logic. Examples taken from this meeting alone include temporal, description, tense, quantum, modal, projective, hybrid, intuitionistic, and linear logics. - eas of application include veri?cation of software and computer systems, ded- tive databases, knowledge representation and its required inference engines, and system diagnosis. The conference brought together researchers interested in all aspects - theoretical foundations, implementation techniques, systems devel- ment and applications - of the mechanization of reasoning with tableaux and related methods.

This book constitutes the refereed proceedings of the 9th International Conference on Database and Expert Systems Applications, DEXA'98, held in Vienna, Austria, in August 1998.
The 81 revised full papers presented were carefully selected from a total of more than 200 submissions. The papers are organized in sections on active databases, object-oriented systems, data engineering, information retrieval, workflow and cooperative systems, spatial and temporal aspects, document management, spatial databases, adaptation and view updates, genetic algorithms, cooperative and distributed environments, interaction and communication, transcation, advanced applications, temporal aspects, oriented systems, partitioning and fragmentation, database queries, data, data warehouses, knowledge discovery and data mining, knowledge extraction, and knowledge base reduction for comprehension and reuse.

This book constitutes the thoroughly refereed post-workshop proceedings of three workshops held in conjunction with the 10th Australian Joint Conference on Artificial Intelligence in Perth, Australia, in December 1997.
The 17 revised full papers presented were carefully reviewed for inclusion in the book. The book is divided into sections on formal methods of agency, reasoning agents, communication and coordination, social interaction, and practical issues for distributed artificial intelligence systems.

This volume brings together the work of researchers from various disciplines where aspects of descriptive, mathematical, computational or design knowledge concerning metaphor and analogy, especially in the context of agents, have emerged. The book originates from an international workshop on Computation for Metaphors, Analogy, and Agents (CMAA), held in Aizu, Japan in April 1998.
The 19 carefully reviewed and revised papers presented together with an introduction by the volume editor are organized into sections on Metaphor and Blending, Embodiment, Interaction, Imitation, Situated Mapping in Space and Time, Algebraic Engineering: Respecting Structure, and a Sea-Change in Viewpoints.

Uncertaintyisanincreasinglyimportantresearchtopicinmanyareasofc- puterscience. Manyformalismsarebeingdeveloped,withmuchinterestatthe theoryleveldirectedatdevelopingabetterunderstandingoftheformalismsand identifyingrelationshipsbetweenformalisms,andatthetechnologyleveldirected atdevelopingsoftwaretoolsforformalismsandapplicationsofformalisms. ThemainEuropeanforumforthesubjectistheEuropeanConferenceon SymbolicandQuantitativeApproachestoReasoningandUncertainty(EC- SQARU). FollowingthesuccessofthepreviousECSQARUconferences,heldin Marseilles(1991),Granada(1993),Fribourg(1995),andBonn(1997),the?fth conferenceintheserieswasheldatUniversityCollegeLondoninJuly1999. ThisvolumecontainspapersacceptedforpresentationatECSQARU'99. In additiontothemainconference,twoworkshopswereheld. The?rstwason DecisionTheoreticandGameTheoreticAgents,chairedbySimonParsonsand MikeWooldridge,andthesecondwasonLogicalandUncertaintyModelsfor InformationSystems,chairedbyFabioCrestaniandMouniaLalmas. Selected papersfromtheworkshopsarealsoincludedintheseproceedings. Weareindebitedtotheprogrammmecommitteefortheire?ortinorganising theprogramme,totheinvitedspeakers,andtothepresentersofthetutorials. Furthermore,wegratefullyacknowledgethecontributionofthemanyreferees whowereinvolvedinthereviewingprocess. Finallywewouldliketothankthe DepartmentofComputerScienceatUniversityCollegeLondonforadminist- tivesupport. ProgrammeCommittee TheprogrammecommitteewaschairedbyAnthonyHunter(UniversityCollege London),andcomprisedDovGabbay(King'sCollegeLondon),FinnJensen (AalborgUniversity),RudolfKruse(UniversityofMagdeburg),SimonParsons (QueenMary,UniversityofLondon)HenriPrade(IRIT,Toulouse),Torsten Schaub(UniversityofPotsdam),andPhilippeSmets(ULB,Bruxelles). Reviewers Theprogrammecommitteeisverygratefulforallthehardworkcontributed bythereviewers. Hopefully,wehavenotmissedanyonefromthefollowing list:BruceD'Ambrosio,FlorenceBannay,SalemBenferhat,PhilippeBesnard, HuguesBersini,ChristianBorgelt,RachelBourne,StefanBrass,LaurenceCholvy, RogerCooke,AdnanDarwiche,YannisDimopoulos,JurgenDix,DidierDubois, UweEgly,LindavanderGaag,JoergGebhardt,SiegfriedGottwald,RolfHaenni, Jean-YvesJa?ray,RadimJirousek,RuthKempson,U?eKjaerulf,FrankK- wonn,AljoschaKlose,JuergKohlas,PaulKrause,GerhardLakemeyer,Mounia Lalmas,JeromeLang,KimG. Larsen,NorbertLehmann,T. Y. Lin,Thomas Linke,KhalidMellouli,JeromeMengin,J. -J. Ch. Meyer,SanjayModgil,Yves Moinard,Sera?nMoral,DetlefNauck,AnnNicholson,PascalNicolas,Dennis VI Preface Nilsson,KristianG. Olesen,RainerPalm,ZdzislawPawlak,VincentRisch,Regis Sabbadin,CamillaSchwind,PrakashP. Shenoy,MilanStudeny,HeikoTimm, HansTompits,MarcoValtorta,andCeesWitteven. April1999 AnthonyHunterandSimonParsons TableofContents Onthedynamicsofdefaultreasoning GrigorisAntoniou...1 Non-monotonicandparaconsistentreasoning:Frombasicentailmentsto plausiblerelations OferArieliandArnonAvron ...11 AcomparisonofsystematicandlocalsearchalgorithmsforregularCNF formulas Ram'onB'ejarandFelipMany'a...22 Query-answeringinprioritizeddefaultlogic FaridBenhammadi,PascalNicolasandTorstenSchaub...32 Updatingdirectedbeliefnetworks BoutheinaBenYaghlaneandKhaledMellouli...43 Inferringcausalexplanations PhilippeBesnardandMarie-OdileCordier...55 Acritiqueofinductivecausation ChristianBorgeltandRudolfKruse...68 Connectinglexicographicwithmaximumentropyentailment RachelA. BourneandSimonParsons...80 Avoidingnon-groundvariables StefanBruning .. andTorstenSchaub...92 Anchoringsymbolstovisiondatabyfuzzylogic SilviaCoradeschiandAlessandroSa?otti...104 Filteringvsrevisionandupdate:Letusdebate! CorineCossartandCatherineTessier...116 Irrelevanceandindependenceaxiomsinquasi-Bayesiantheory FabioG. Cozman...128 Assessingthevalueofacandidate:Aqualitativepossibilisticapproach DidierDubois,MichelGrabischandHenriPrade...137 VIII TableofContents Learningdefaulttheories B'eatriceDuvalandPascalNicolas...148 Knowledgerepresentationforinductivelearning PeterA. Flach...160 Handling inconsistency e?ciently in the incremental construction of strati?ed beliefbases EricGr'egoire...168 Roughknowledgediscoveryandapplications J. W. GuanandD. A. Bell...179 GradientdescenttrainingofBayesiannetworks FinnV. Jensen...190 Opendefaulttheoriesovercloseddomains:Anextendedabstract MichaelKaminski...201 Shopboteconomics Je?reyO. KephartandAmyR. Greenwald...208 OptimizedalgorithmforlearningBayesiannetworkfromdata F'ediaKhalfallahandKhaledMellouli...221 Mergingwithintegrityconstraints S'ebastienKoniecznyandRam'onPinoP'erez...233 Boolean-likeinterpretationofSugenointegral IvanKramosil...245 AnalternativetooutwardpropagationforDempster-Shaferbelieffunctions NorbertLehmannandRolfHaenni...256 Onbottom-uppre-processingtechniquesforautomateddefaultreasoning ThomasLinkeandTorstenSchaub...2 68 Probabilisitclogicprogrammingundermaximumentropy ThomasLukasiewiczandGabrieleKern-Isberner...279 Lazypropagationandindependenceofcausalin?uence AndersL. MadsenandBruceD'Ambrosio...293 TableofContents IX AMonteCarloalgorithmforcombiningDempster-Shaferbeliefbasedon approximatepre-computation Seraf'?nMoralandAntonioSalmer'on...305 Anextensionofalinguisticnegationmodelallowingustodenynuanced propertycombinations DanielPacholczyk...316 Argumentationandqualitativedecisionmaking SimonParsonsandShawGreen...328 Handling di?erent forms of uncertainty in regression analysis: A fuzzy belief structureapproach SimonPetit-RenaudandThierryDenoeux...340 State recognition in discrete dynamical systems using Petri nets and evidence theory Mich'eleRombaut,ImanJarkassandThierryDenoeux...352 Robotnavigationandmapbuildingwiththeeventcalculus MurrayShanahanandMarkWitkowski...362 Informationfusioninthecontextofstockindexprediction StefanSiekmann,J..orgGebhardtandRudolfKruse...363 Defeasiblegoals LeendertvanderTorre...374 Logicaldeductionusingthelocalcomputationframework NicWilsonandJ'eromeMengin...386 AuthorIndex...397 OntheDynamicsofDefaultReasoning Grigoris Antoniou Gri?thUniversity,QLD4111,Australia UniversityofMacedonia,Thessaloniki,Greece ga@cit. gu. edu. au Abstract. Defaultlogicisaprominentrigorousmethodforreasoning withincompleteinformationbasedonassumptions. Itisastaticreas- ingapproach,inthesensethatitdoesn'treasonaboutchangesandtheir consequences. Ontheotherhand,itsnonmonotonicbehaviourappears whenchangestoadefaulttheoryaremade.

This two-volume set constitutes the refereed proceedings of the 11th International Conference on Industrial and Engineering Applications of Artificial Intelligence and Expert Systems, IEA/AIE-98, held in Benicassim, Castellon, Spain, in June 1998.The two volumes present a total of 187 revised full papers selected from 291 submissions. In accordance with the conference, the books are devoted to new methodologies, knowledge modeling and hybrid techniques. The papers explore applications from virtually all subareas of AI including knowledge-based systems, fuzzyness and uncertainty, formal reasoning, neural information processing, multiagent systems, perception, robotics, natural language processing, machine learning, supervision and control systems, etc..

Reasoning with incomplete information constitutes a major challenge for any intelligent system. In fact, we expect such systems not to become paralyzed by missing information but rather to arrive at plausible results by bridging the gaps in the information available.
A versatile way of reasoning in the absence of information is to reason by default. This book aims at providing formal and practical means for automating reasoning with incomplete information by starting from the approach taken by the framework of default logic. For this endeavor, a bridge is spanned between formal semantics, over systems for default reasoning, to efficient implementation.

A special year like 1999 invites one to draw a balance of what has been achieved in the roughly 30 years of research and development in knowledge based systems (still abbreviated as XPS following the older term "expert systems") and to take a look at th what the future may hold. For the 5 German conference on knowledge-based systems we therefore asked current and former speakers of the four working groups (FG's) in the subdivision of knowledge-based systems (FA 1.5) of the German association of Informatics (GI) to present a survey of and future prospects for their respective fields: knowledge engineering, diagnosis, configuration, and case-based reasoning. An additional 14 technical papers deal with current topics in knowledge-based systems with an equal emphasis on methods and applications. They are selected from more than 50 papers accepted in the 4 parallel workshops of XPS-99: a) Knowledge Management, Organizational Memory and Reuse, b) various fields of applications, c) the traditional PuK Workshop (planning and configuration), and d) the GWCBR (German workshop on case-based reasoning). The other papers presented at these workshops are not included in this volume but are available as internal reports of Wurzburg university together with the exhibition guide that emphasizing tool support for building knowledge based systems."

The journey towards the autonomous enterprise has begun; there are already companies operating in a highly automated way. Every corporate decision-maker will need to understand the opportunities and risks that the autonomous enterprise presents, to learn how best to navigate the shifting competitive landscape on their journey of change. This book is your guide to this innovation, presenting the concepts in real world contexts by covering the art of the possible today and providing glimpses into the future of business.

The first International Conference on Intelligent Tutoring Systems (ITS) was held ten years ago in Montreal (ITS '88). It was so well received by the international community that the organizers decided to do it again in Montreal four years later, in 1992, and then again in 1996. ITS '98 differs from the previous ones in that this is the first time the conference has been held outside of Montreal, and it's only been two years (not four) since the last one. One interesting aspect of the ITS conferences is that they are not explicitly bound to some organization (e.g., IEEE or AACE). Rather, the founder of these conferences, Claude Frasson, started them as a means to congregate researchers actively involved in the ITS field and provide a forum for presentation and debate of the most currently challenging issues. Thus the unifying theme is science. This year's "hot topics" differ from those in the earlier ITS conferences as they reflect ever changing trends in ITS research. A few of the issues being examined at ITS '98 include: Web based tutoring systems, deploying ITS in the real world, tutoring and authoring tools, architectures, and knowledge structure and representation.

This book provides a systematic in-depth investigation of a class of multiple-context assumption-based multiagent reasoning problems, typical, e.g., for distributed planning, scheduling, and control. First, logical and architectural foundations are provided to construct the two systems XFRMS and MXFRMS allowing the development of more complex utilities. Then the technology developed for XFRMS in the single-agent case is further enhanced to the multiagent situation. The multiagent RMS MXFRMS provides, along with XFRMS, a solid foundation for building more complex utilities, such as plan or schedule maintenance systems. In the final chapter, the software engineer confronted with the task of building a practicable multiagent system can find a discussion of potential uses and future extensions of the systems provided.

This book constitutes the refereed proceedings of the Second International Workshop on Cooperative Information Systems, CIA'98, held in cognition with Agents World in July 1998 in Paris.
The book presents nine invited contributions together with 14 revised full papers selected from a total of 54 submissions. The book is divided in parts on systems and applications; issues of design, querying, and communication; rational cooperation and electronic commerce; adaptive and collaborative information gathering; and mobile information agents in the internet.

This book contains the refereed proceedings of the 8th International Conference on Database and Expert Systems Applications, DEXA '97, held in Toulouse, France, September 1997. The 62 revised full papers presented in the book, together with three invited contributions, were selected from a total of 159 submissions. The papers are organized in sections on modeling, object-oriented databases, active and temporal aspects, images, integrity constraints, multimedia databases, deductive databases and knowledge-based systems, allocation concepts, data interchange, digital libraries, transaction concepts, learning issues, optimization and performance, query languages, maintenance, federated databases, uncertainty handling and qualitative reasoning, and software engineering and reusable software.

This book constitutes the refereed proceedings of the Second International Conference on Case-Based Reasoning, ICCBR-97, held in Providence, RI, USA, in July 1997.
The volume presents 39 revised full scientific papers selected from a total of 102 submissions; also included are 20 revised application papers. Among the topics covered are representation and formalization, indexing and retrieval, adaptation, learning, integrated approaches, creative reasoning, CBR and uncertainty. This collection of papers is a comprehensive documentation of the state of the art in CBR research and development.

This two-volume set constitutes the refereed proceedings of the 11th International Conference on Industrial and Engineering Applications of Artificial Intelligence and Expert Systems, IEA/AIE-98, held in Benicassim, Castellon, Spain, in June 1998.The two volumes present a total of 187 revised full papers selected from 291 submissions. In accordance with the conference, the books are devoted to new methodologies, knowledge modeling and hybrid techniques. The papers explore applications from virtually all subareas of AI including knowledge-based systems, fuzzyness and uncertainty, formal reasoning, neural information processing, multiagent systems, perception, robotics, natural language processing, machine learning, supervision and control systems, etc..

This book constitutes the refereed proceedings of the International Conference on Analytic Tableaux and Related Methods, TABLEAUX'97, held in Pont-a-Mousson, France, in May 1997.
The volume presents 22 revised full papers selected from a total of 49 submissions. Also included are two invited papers and two system descriptions. The volume covers the whole spectrum of tableaux-based theorem proving and its applications including theoretical foundations, methodological issues, implementation techniques, and system development. Besides classical logics, among the logics dealt with are modal, intuitionistic, many-valued, and temporal logic.

This book constitutes the refereed proceedings of the First International Joint Conference on Qualitative and Quantitative Practical Reasoning, ECSQARU-FAPR'97, held in Bad Honnef, Germany, in June 1997.
The volume presents 33 revised full papers carefully selected for inclusion in the book by the program committee as well as 12 invited contributions. Among the various aspects of human practical reasoning addressed in the papers are nonmonotonic logics, default reasoning, modal logics, belief function theory, Bayesian networks, fuzzy logic, possibility theory, inference algorithms, dynamic reasoning with partial models, and user modeling approaches.