Encyclopedia of GIS

2017 Edition
| Editors: Shashi Shekhar, Hui Xiong, Xun Zhou

Geospatial Semantic Web: Personalization

  • Pragya Agarwal
Reference work entry
DOI: https://doi.org/10.1007/978-3-319-17885-1_517

Synonyms

Definition

What Is the Semantic Web?

The basic principle underlying the semantic web is the idea of having data defined and linked in such a way that it can be used for more effective discovery, automation, integration, and reuse across various applications (Berners-Lee et al. 2001; Fensel 2002). With the idea of a semantic web in which machines can understand, process and reason about resources to provide better and more comfortable support for humans in interacting with the World Wide Web, the question of personalizing the interaction with web content to user needs and profile is of prime significance.

What Is the Geospatial Semantic Web?

Geographic data has meanings associated with it and semantic annotations of geographic data will allow for better retrieval and more effective use of geospatial services. Recent developments in the semantic web have great potential for the geospatial community, specifically because the focus on the incorporation of data semantics will lead to a better retrieval and more reliable integration methods by tapping into the semantics during the search process on the web. There is a distinct move away from structure and syntax in the geospatial community, accompanied by an increased awareness that semantics is the backbone for a successful ontology to enable translation of data from different resources and users. Ontologies are being increasingly used in the geospatial community to enable interoperability between multiple resources, systems, services and semantic conceptualizations. The most commonly used definition of an ontology is that it is a “specification of a conceptualization” (Gruber 1993). However, the basic semantic web and the technological developments are not targeted to the needs of the geospatial community as spatial data has its own specific needs associated with it. A geospatial domain is characterized by vagueness, especially in the semantic disambiguation of the concepts in the domain, which makes defining universally accepted geoontology an onerous task (Agarwal 2005). This is compounded by the lack of appropriate methods and techniques where the individual semantic conceptualizations can be captured and compared to each other. The idea of a more focused “Geospatial Semantic Web” has been recognized as a research priority within UCGIS (University Consortium for Geographic Information Science) initiatives (Fonseca and Sheth 2002). Egenhofer (2002) identified the need to support queries based on meanings and better definition of spatial terms across a number of disciplines, and the need to integrate multiple terminological ontologies as a backbone for an effective geospatial semantic web (GSW).

What Is Personalization?

Nowadays, most large-scale applications are planned and designed for a large variety of end users. Nevertheless, the traditional “one-size-fits-all” approach turns out to be outdated and inappropriate to meet with heterogeneous user needs. “Personalization” is the key word and the specific challenge is to find out what the users are looking for in order to understand how to offer them tailored information. Therefore, personalization of information means to deal with information more efficiently and effectively, performing complex tasks within a reasonable time, in a “user-friendly” environment. The development of any personalization techniques and adaptive information retrieval systems should deal with the users’ understanding of domain examined by knowledge capture approaches.

A “closed world” paradigm in which individual preferences and goals are not considered, in which the system requirements are deterministic, and not adapted to changing needs of the community, has meant that the system design and operation work on a fixed set of resources which are normally known to the system designers at design time. On the other hand, personalization facilitates a move away from this deterministic approach that traditional information system design frameworks operate under. The individual requirements and context that the user operates under has to be taken into account to move from a closed-world setting. knowLedge-enhanced web services are normally driven by some description of the world which is encoded in the system in the form of an ontology defined by knowledge engineers. The users’ conceptualization of the world may also differ, sometimes significantly, from the conceptualization encoded in the system. If not taken into account, the discrepancies between a user’s and a system’s conceptualization may lead to the user’s confusion and frustration when utilizing semantic web services, which, in turn, can make these services less popular. Semantic personalisation of web-based services is required to exploit user intentions and perspectives and tailor the information accordingly.

Personalization can be defined as a process of gathering user-information during interaction with the user, which is then used to deliver appropriate content and services, tailored to the user’s needs. Personalization is a process by which it is possible to give the user optimal support in accessing, retrieving, and storing information, where solutions are built so as to fit the preferences, the characteristics and the taste of the individual. Personalization of information systems aims at developing systems that can autonomously interoperate-either with humans or with other systems-adapting their processing and its outcome to specific requests. Personalized information systems aim to make user needs the center of the interaction process and to optimize user access to the relevant and required information according to individual needs, profiles and particular contexts. This result can be achieved only by exploiting machine-interpretable semantic information, e.g., about the possible resources, about the user him/herself, about the context, about the goal of the interaction. Personalization is realized by a reasoning process applied to the semantic information, which can be carried out in many different ways depending on the specific task.

Although, conventionally Personalization is taken to mean directing the system design to user needs and profiles, it can also mean to adapt according to local browser or regional contexts. The individual requirements of the user are to be taken into account in such different dimensions. Some of these individual user requirements will include the current task, the goal of the user, the context in which the user is requesting the information, the previous information requests or interactions of the user, the working process s/he is involved in, knowledge of the user (an expert will be satisfied by information which is not suitable for a layman), the device s/he is using to display the information, the bandwidth and availability of the communication channel, the abilities/disabilities/handicaps of the user, and the time constraint of the user (whether s/he is under time pressure, or is just browsing some information).

Personalization of Semantic Web

Personalized semantic web is the concept of the semantic web in which machines are enabled to understand the meaning of information and thereby provide better support to individuals in carrying out their tasks, and is aimed at improving the user’s experience of a web-based service. In particular, applications that can retrieve, process and present information in enhanced user-adapted ways are interesting.

Many attempts have been made to apply personalization techniques to the World Wide Web as a natural extension of work on hypertext and hypermedia; however, the web is an information space thought for human to human communication, while personalization requires software systems to take part to the interaction and help (Dolog et al. 2003). Such systems require knowledge to be expressed in a machine-interpretable format, which in the web is not available. The development of languages for expressing information in a machine-processable form is characteristic of the semantic web initiative (Berners-Lee 1998). Over this knowledge layer, the use of inferencing mechanisms is envisioned as a fundamental means for performing a content-aware navigation, producing an overall behavior that is closer to the user’s intuition and desire. This is the reason that the semantic web is the most appropriate environment for realizing personalization. In other words, the semantic web is deeply connected to the idea of personalization in its very nature.

What Is Personalization of GSW?

Users’ preferences, expectations, goals and tasks differ while using the web for geographic information. Moreover, people form different conceptual models of the world and these models dynamically change over time. In addition, meanings are crucial in distinction of geographic information and people constantly assign meanings to real world objects, while categorizing them as they interact with the world around them. Agarwal (2005) discusses in detail the problems associated with ontology development in the geospatial domain primarily due to semantic ambiguities. The knowledge-enhanced web services are normally driven by some description of the world which is encoded in the system in the form of an ontology defined by knowledge engineers. The user’s conceptualization of the world may differ, sometimes significantly, from the conceptualization encoded in the system. If not taken into account, the discrepancies between a user’s and a system’s conceptualization may lead to the user’s confusion and frustration when utilizing the web-based geospatial services, which, in turn, can make these services less popular. Indeterminacy and ambiguity in meanings are key issues in the development of ontologies in the geographic domain (Agarwal 2005; Bennett 2001). Empirical results show that individual conceptualizations are characterized by semantic heterogeneity (Agarwal 2004).

With multiple user conceptualizations, efforts towards a reliable geospatial semantic web, therefore, require personalization where user diversity can be incorporated and targeted to multiple needs, semantics and conceptualizations of the real world. Egenhofer (2002) identified the need to support queries based on meanings and better definition of spatial terms across a number of disciplines, and the need to integrate multiple terminological ontologies as a backbone for an effective GSW. The success of a standardized geoontology for the semantic web will be determined by the level of acceptance by the users of the services, both expert and naïve, and the level to which the basic geoontology is semantically compatible with the users’ conceptuali-zations.

In this way, personalization is crucial for acknowledging and addressing individual user differences and needs and providing user-friendly user access to geospatial information on the web. To achieve a GSW will require both syntactic and semantic interoperability of resources, and therefore the personalization efforts are essential in making the GSW a reality.

Historical Background

GSW is a new initiative and many concrete applications of it are yet to be seen. Personalization of GSW is a novel concept and has no historical precedents.

Scientific Fundamentals

Many research disciplines have contributed to web personalization research, for example, hypertext research has studied personalization in the area of so-called adaptive hypertext systems, collaborative filtering research has investigated recommender systems, artificial intelligence techniques have been widely used to cluster web data, usage data, and user data, reasoning and uncertainty management has been adopted to draw conclusions on appropriate system behavior, etc. Dolog et al. (2003). However, in most systems, there are no mechanisms to capture the interaction and context of the user. There is an urgent need to include the people as an axis in the design, development, and deployment of semantically enriched services, especially for personalization of the GSW where individual user semantics vary. In addition, computational models are needed that can process the different terminological and semantic ontologies and process the semantic incompatibilities between users and the expert’s geoontology.

Resolving the discrepancy between psychological user variables and physical system variables in the area of geospatial services goes beyond the user-interface level. Rather than a closed view of the world, the personalization efforts for geospatial services design will ensure that the different perspectives and semantic conceptualizations of the real world are maintained as “open”. The underlying principle for the methodology adopted by Agarwal et al. (2005) and Huang et al. (2005) is that such geospatial services and applications could benefit from personalization efforts where semantic variations and conceptualizations are captured and aligned, and differences with the core expert ontology identified and formalized to provide the basis for user-supported access to such location-based information.

This will allow, first, for the development of systems that allow personalization by incorporating user models and diversity and second, as a means to test any core ontologies that are developed as the basis for a geospatial services against user conceptualizations for discrepancies and thereby evaluate its reliability as a standard, reusable ontology. Moreover, the personalization approach allows flexibility and the possibility of using the user models to enrich the available information resources with shared semantics instead of relying on fixed ontologies available to the developers at the design stage. Using this approach, the research towards specification of well-defined standards and ontologies for interoperability in the geospatial domain can be enhanced and personalized to provide extendibility, flexibility, interoperability, and reusability. Automating the mapping of multiple conceptualizations and personalization of web-based services will also facilitate pervasive computing in mobile services and enable more effective use of mobile GIS services.

Key Applications

The availability of geospatial knowledge resources on the web enables members of the public to take advantage of trusted knowledge built by domain experts, e.g., for planning travel routes and for accessing weather information. Geospatial services and systems are also unique in the way that they use data, which are related to locations in space and time, and that the processing of the data with respect to these spatial locations is possible. People’s input to geospatial tools have a spatiotemporal context. One can ask “where is a certain object” or “where are all objects with certain properties” at a given time when trying to find the nearest health services for the elderly; or one can ask “what are the properties of a certain area in space (as well as time)” when trying to ascertain the suitability of an environment (for example, crime rates) while renting or buying a property. Users access the web services with different goals and often; these services require integration of the various different resources to provide a comprehensive result for the user search for their specific requirements. For example, in a “what is in my backyard” service provided by the Environment Agency (EA) in the UK, members of the public can see what pollutants may be scattered across their neighborhood. End-users will have their own contexts of use: property evaluation, ecology, etc. and for a member of the public, a general interest (based on a topographic view of different areas in the city). Each could potentially view the data provided by the others but form their own conceptual understanding of the location-based information. Personalization efforts will make it possible to capture and formalize context and thereby provide context-relevant information.

Future Directions

There is no common language for expressing adaptive functionality, hence these systems are difficult to compare and analyze, also suffering to a great extent from lack of reusability-which in fact is the key capability for successful personalization functionality for the GSW. To deal with a diverse user population having different preferences, goals, understanding of tasks and conceptual models, existing design paradigms in geospatial services will have to be redefined. Furthermore, new diagnostic techniques and models are needed to capture the long-term development of users’ capabilities, the dynamics of users’ goals and conceptual understanding, the uncertainty and inconsistency of naive users’ conceptualizations, and so on. The ambitious target is to offer manageable, extendible and standardized infrastructure for complementing and collaborating applications tailored to the needs of individual users. Traditional personalization and adaptation architectures were suited to deal with closed-world assumption, where user modeling methods, such as overlay, bug library, constraint-based modeling and other marked discrepancies in a user and expert’s semantics as erroneous, and often called them misconceptions. New approaches for open-world user modeling that facilitate elicitation of extended models of users are needed to deal with the dynamics of a user’s conceptualization. Similarly, methods that acknowledge semantic discrepancies and heterogeneity are required for effectively personalizing the web-based services for the geospatial community. Adaptation and personalization methods are not developed to address the temporality inherent in geospatial information and therefore, any specific personalization efforts for the GSW will have to be modified to suit this need. Without the benefit of deeper semantic or ontological knowledge about the underlying domain, personalization systems cannot handle heterogeneous and complex objects based on their properties and relationships. Nor can these systems possess the ability to automatically explain or reason about the user models or user recommendations. This realization points to an important research focus that can combine the strengths of web mining with semantic or ontological knowledge. Development of automated reasoning tools to detect mismatches and discrepancies between the user and the expert ontology forming the backbone for the web-based resources will be a step forward.

Cross-References

References

  1. Agarwal P (2004) Contested nature of place. In: Egenhofer M, Freska C, Harvey M (eds) 3rd international conference GIScience (2004). Lecture notes in computer science, vol 3234. Springer, Berlin, pp 1–21Google Scholar
  2. Agarwal P (2005) Ontological considerations in GIScience. Int J Geogr Inf Sci 19:501–535CrossRefGoogle Scholar
  3. Agarwal P, Huang Y, Dimitrova V (2005) Formal approach to reconciliation of individual ontologies for personalization of geospatial semantic web. In: Proceedings of the first international conference on geospatial semantics, Mexico City, Nov 2005. Lecture notes in computer science, vol 3799. Springer, Berlin, pp 202–219Google Scholar
  4. Bennett B (2001) What is a forest? On the vagueness of certain geographic concepts. Topoi 20: 189–201CrossRefGoogle Scholar
  5. Berners-Lee T (1998) The semantic web road map. World Wide Web consortium W3C draft. http://www.w3.org/DesignIssues/Semantic
  6. Berners-Lee T, Hendler J, Lassila O (2001) The semantic web. Sci Am 284(5):28–37CrossRefGoogle Scholar
  7. Dolog P, Henze N, Nejdl W, Sintek M (2003) Towards the adaptive semantic web. In: Proceedings of the 1st workshop on principles and practice of semantic web reasoning (PPSWR’03), co-located with international conference on logic programming (ICLP’03), Mumbai, 8 Dec 2003Google Scholar
  8. Egenhofer MJ (2002) Toward the semantic geospatial web. In: Proceedings of the tenth ACM international symposium on advances in geographic information systems, McLean, 8–9 Nov 2002Google Scholar
  9. Fensel D, Wahlster W, Lieberman L, Hendler J (2002) Spinning the semantic web: bringing the world wide web to its full potential. MIT, CambridgeGoogle Scholar
  10. Fonseca F, Sheth A (2002) The geospatial semantic web. UCGIS white paper. http://www.ucgis4.org/priorities/research/2002researchagenda.htm
  11. Gruber TR (1993) A translation approach to portable ontologies. Knowl Acquis 5:199–220CrossRefGoogle Scholar
  12. Huang Y, Dimitrova V, Agarwal P (2005) Detecting mismatches between a user’s and a system’s conceptualizations to empower user modelling for the Semantic Web. In: Proceedings of the workshop on personalization in the semantic web, user modelling conference, Edinburgh, 24–30 July 2005Google Scholar

Recommended Reading

  1. Antoniou G, van Harmelen F (2004) A semantic web primer. MIT, CambridgeGoogle Scholar
  2. Brusilovsky P (2001) Adaptive hypermedia. User Model User Adapt Interact 11:87–110MATHCrossRefGoogle Scholar
  3. Cardoso J (2007) Semantic web services: theory, tools and applications. Idea GroupCrossRefGoogle Scholar
  4. Cardoso J, Sheth A (2006) Semantic web services, processes and applications. Springer, New YorkMATHCrossRefGoogle Scholar
  5. Davies J, Fensel D, van Harmelen F (2003) Towards the semantic web: ontology-driven knowledge management. Wiley, Chichester/HobokenGoogle Scholar
  6. Doan A, Madhavan J, Domingos P, Halvey A (2002) Learning to map between ontologies on the semantic web. In: Proceedings of WWW’02, Hanau, 7–11 May 2002Google Scholar
  7. Dolog P, Henze N, Nejdl W, Sintek M (2004) Personalization in distributed e-learning environments. In: Proceedings of 13th world wide web conference, New York, 17–22 May 2004Google Scholar
  8. Hendler J, Berners-Lee T, Miller E (2002) Integrating applications on the semantic web (English version). J Inst Electr Eng Jpn 122:676–680CrossRefGoogle Scholar
  9. Shadbolt N, Berners-Lee T, Hall W (2006) The semantic web revisited. IEEE Intell Syst 21:96–101CrossRefGoogle Scholar
  10. Staab S, Studer R (2004) Handbook on ontologies. Springer, HeidelbergMATHCrossRefGoogle Scholar

Copyright information

© Springer International Publishing AG 2017

Authors and Affiliations

  • Pragya Agarwal
    • 1
  1. 1.Department of Geomatic EngineeringUniversity College LondonLondonUK