Angela Schwering

Approaches to Semantic Similarity Measurement for Geo-Spatial Data: A Survey

Semantic similarity is central for the functioning of semantically enabled processing of geospatial data. It is used to measure the degree of potential semantic interoperability between data or different geographic information systems (GIS). Similarity is essential for dealing with vague data queries, vague concepts or natural language and is the basis for semantic information retrieval and integration. The choice of similarity measurement influences strongly the conceptual design and the functionality of a GIS. The goal of this article is to provide a survey presentation on theories of semantic similarity measurement and review how these approaches – originally developed as psychological models to explain human similarity judgment – can be used in geographic information science. According to their knowledge representation and notion of similarity we classify existing similarity measures in geometric, feature, network, alignment and transformational models. The article reviews each of these models and outlines its notion of similarity and metric properties. Afterwards, we evaluate the semantic similarity models with respect to the requirements for semantic similarity measurement between geospatial data. The article concludes by comparing the similarity measures and giving general advice how to choose an appropriate semantic similarity measure. Advantages and disadvantages point to their suitability for different tasks.

Hybrid model for semantic similarity measurement

Expressive knowledge representations with flexible semantic similarity measures are central for the functioning of semantic information retrieval, information integration, matchmaking etc. Existing knowledge representations provide no or not sufficient support to model the scope of properties. While properties in feature- and geometric models always refer to the whole concept, structured representations such as the alignment model provide a limited support for scope by assigning properties to objects which are part of the whole entity. Network models do not support properties at all. In this paper we propose a hybrid model: a structured knowledge representation combining the relational structure of semantic nets with property-based description of feature- or geometric models. It supports to model properties—features or dimensions—and their scope by taxonomic or non-taxonomic relations between a concept and its properties. The similarity measure computes the similarity in consideration of the scope of each property.

Measuring semantic similarity between geospatial conceptual regions

Determining the grade of semantic similarity between geospatial concepts is the basis for evaluating semantic interoperability of geographic information services and their users. Geometrical models, such as conceptual spaces, offer one way of representing geospatial concepts, which are modelled as n-dimensional regions. Previous approaches have suggested to measure semantic similarity between concepts based on their approximation by single points. This paper presents a way to measure semantic similarity between conceptual regions—leading to more accurate results. In addition, it allows for asymmetries by measuring directed similarities. Examples from the geospatial domain illustrate the similarity measure and demonstrate its plausibility.

Spatial relations for semantic similarity measurement

Measuring semantic similarity among concepts is the core method for assessing the degree of semantic interoperability within and between ontologies. In this paper, we propose to extend current semantic similarity measures by accounting for the spatial relations between different geospatial concepts. Such integration of spatial relations, in particular topologic and metric relations, leads to an enhanced accuracy of semantic similarity measurements. For the formal treatment of similarity the theory of conceptual vector spaces—sets of quality dimensions with a geometric or topologic structure for one or more domains—is utilized. These spaces allow for the measurement of semantic distances between concepts. A case study from the geospatial domain using Ordnance Surveys MasterMap is used to demonstrate the usefulness and plausibility of the approach.

Syntactic principles of heuristic-driven theory projection

Analogy making is a central construct in human cognition and plays an important role to explain cognitive abilities. While various psychologically or neurally inspired theories for analogical reasoning have been proposed, there is a lack of a logical foundation for analogical reasoning in artificial intelligence and cognitive science. We aim to close this gap and propose heuristic-driven theory projection (HDTP), a mathematically sound framework for analogy making. HDTP represents knowledge about the source and the target domain as first-order logic theories and compares them for structural commonalities using anti-unification. The paper provides an overview of the syntactic principles of HDTP, explains all phases of analogy making at a formal level, and illustrates these phases with examples.

Semantic Similarity Measurement and Geospatial Applications

With the increasing amount of geographic information available on the Internet, searching, browsing, and organizing such information has become a major challenge within the field of Geographic Information Science (GIScience). As all information is ultimately for and from human beings, the methodologies applied to retrieve and organize this information should correlate with human similarity judgments. Semantic similarity measurement, which originated in psychology, is a methodology fulfilling this requirement and supporting geographic information retrieval. The following special issue presents work on semantic similarity measurement from different perspectives, including cognitive science, information retrieval, and ontology engineering, with a focus on applications in GIScience. It originated in the Workshop on Semantic Similarity Measurement and Geospatial Applications held in conjunction with COSIT 2007, the International Conference on Spatial Information Theory (http://www.cosit.info/). A substantial part of the workshop contributions addressed the need for similarity measurement in geographic information retrieval, including applications in web service discovery, knowledge management, and emergency scenarios. The call for papers to this issue was based on these workshop contributions, discussions, and results (http://musil.uni-muenster.de), but open to any submissions on the role of semantic similarity in GIScience. Eleven papers were submitted and then

Restricted higher-order anti-unification for analogy making

Anti-unification has often be used as a tool for analogy making. But while first-order anti-unification is too simple for many applications, general higher-order anti-unification is too complex and leads into theoretical difficulties. In this paper we present a restricted framework for higher-order substitutions and show that anti-unification is well-defined in this setting. A complexity measure for generalizations can be introduced in a quite natural way, which allows for selecting preferred generalizations. An algorithm for computing such generalizations is presented and the utility of complexity for anti-unifying sets of terms is discussed by an extended example.

Evaluation of a semantic similarity measure for natural language spatial relations

Consistent and flawless communication between humans and machines is the precondition for a computer to process instructions correctly. While machines use well-defined languages and formal rules to process information, humans prefer natural language expressions with vague semantics. Similarity comparisons are central to the human way of thinking: we use similarity for reasoning on new information or new situations by comparing them to knowledge gained from similar experiences in the past. It is necessary to overcome the differences in representing and processing information to avoid communication errors and computation failures.We introduce an approach to formalize the semantics of natural language spatial relations and specify it in a computational model which allows for similarity comparisons. This paper describes an experiment that investigates human similarity perception between spatial relations and compares it to the similarity determined by the our semantic similarity measure.

SketchMapia: Qualitative Representations for the Alignment of Sketch and Metric Maps

Abstract: More and more private citizens collect and publish environmental data via web-based geographic information systems. These systems face two challenges: The user interface must be intuitive and the processing of geographic information must account for cognitive impact. We propose to use sketch maps as the medium for interaction, because they reflect a persons spatial knowledge. Information from sketch maps is distorted, schematized, incomplete, and generalized and metric maps are not. This article employs qualitative representations for the alignment of sketch and metric maps. We suggest a set of cognitively oriented aspects in sketch maps stably computed by people and evaluate qualitative representations to formalize these aspects. This allows us to align and integrate geographic information from sketch maps.

An Empirical Study on Relevant Aspects for Sketch Map Alignment

Sketch maps are drawn from memories and they are in general schematized and distorted. However, the schematizations and distortions are not random. They are a consequence during the cognitive process of perceiving, memorizing, and producing spatial layout. This paper describes an empirical study to investigate the impact of distortions on similarity perception. The study is designed as a human-subjects experiment of similarity ranking with two scenarios. Subjects were presented with 45 sketch maps and one reference map in each scenario; they were asked to rank the sketch maps according to their similarities with the reference map. The results of the experiment are used to develop a cognitively motivated alignment strategy for computer-based comparison of sketch maps and metric maps.