Font Size: a A A

Research On Theories And Methods Of Personalized Map Service

Posted on:2012-06-25Degree:DoctorType:Dissertation
Country:ChinaCandidate:Z H WuFull Text:PDF
GTID:1110330371462485Subject:Cartography and Geographic Information System
Abstract/Summary:PDF Full Text Request
Thanks to the development of new spatial information technology, map service has changed in many aspects, such as popularization of users, real time updating of geospatial database, diversification of interactive device, personalized idea of service. So, personalized map service (PMS) comes forth to be a challenged project in cartography study. A new mode of PMS based on adaptable customization and adaptivity is brought out, and detailed researches and practices are done about its theories and methods in this thesis. The main works and creations are as the following:1. Analysis of the background and current research of PMS and its related fields. Personal service theories, techniques and methods in PMS and its related fields are reviewed, especially, current research of PMS in the aspects of system implement, theory framework, technique framework, knowledge engineering, context modeling, evaluation are analyzed, and the main problems are pointed out.2. Study on PMS theory and technique framework, including concept, theory framework and technique framework. In the study of concept, the basic concept, characters and contents of PMS are explicated, the key issues of PMS are brought out and analyzed, and some special issues in web PMS and mobile PMS are pointed out. In the study of theory framework, the basic theories and methods are summarized and organized, and several theories and their application in PMS are detailed introduced. In the study of technique framework, the data stream of PMS system is analyzed, and several main techniques in each step of the framework are introduced.3. Study on theories and methods of knowledge engineering in PMS. Theories and methods of knowledge engineering in PMS are studied systematically, in order to provide theories and methods for this knowledge-based system. The following matters of PMS knowledge engineering are pointed out and studied: contents and characteristics of PMS knowledge; PMS knowledge acquisition, representation, inference and interpretation mechanism; and PMS knowledge database establishment and maintenance.4. Study on experiments and algorithms for acquisition of PMS templates and rules. The following four typical experiments are detailedly introduced: rules acquisition of cartography visualization adapting to devices, rules acquisition of cartography visualization adapting to user visual perception and cognition, user interest acquisition based on browse behavior, rules acquisition based on customization behavior. Algorithms used includes statistic methods, ID3 algorithm in decision tree, interestingness estimate algorithm based on browse behavior, and Apriori association rule mining algorithm considering user interestingness. And the correlative experiments are listed. These experiments and algorithms are of great importance for acquisition of PMS templates and rules.5. Study on multi-dimensional context modeling in PMS. The contents and representation of PMS classified context models are discussed. A multivariate function and characters of integrated context model are brought forward, and its clustering strategy and related algorithms are improved. The concept of subspace is introduced into context clustering; the weight vector of attribute set is calculated based on entropy with ID3 algorithm; similarity measurement based on weighted distance between corresponding subspaces is done, and then subspaces clusters of contexts are created based on OPTICS density clustering algorithm, and then integrated context models are constructed after contexts are endued with rules. In order to solve the following two problems in context model matching: the difficulty to choose out the optimal one from several models with similarities of faint variation, and the influence of time-serial of contexts on model matching, a model of context model matching method based on probability and time-serial is brought out. The dynamic updating principle of multi-dimensional context model is studied, and the dynamic updating algorithms of its function model and user interest model are described.6. Study on evaluation of PMS system. A PMS system evaluation framework integrating layered evaluation and heuristic evaluation is put forward. The methods and techniques for PMS system evaluation including adaptive system evaluation methods and cartography cognition experiment methods are discussed, and their applicability for different steps and different layers is given out. The evaluation criteria is discussed, and based on which, an evaluation index system is constructed. At last, the relationship of evaluation and design of PMS system is emphasized.7. Design and implementation of a PMS prototype system. Based on study of PMS theories and techniques, a prototype system---personalized tourist web map service system is designed and implemented using ASP.NET Ajax. The technique framework and function framework of prototype system are constructed. The templates of user interface, map symbol, color design, navigation function and data recommendation are designed for tourist theme. Some adaptive functions, real-time personal customization functions and online personal customization—outline processing functions are designed and implemented. A primary evaluation of this prototype system is carried out.
Keywords/Search Tags:Personalized Map Service, Customization, Adaptivity, Multi-dimensional Context, Theory and Technique Framework, Knowledge Engineering, Templates and Rules, Experiment, ID3 Algorithm, Interestingness Estimate Algorithm, Apriori Algorithm, Context Modeling
PDF Full Text Request
Related items