National Spatial Data Infrastructure

Managing data heterogeneity is an ardous task to achieve
syntactic, structural and semantic homogeneity that is essential
for proper usage of data. This article presents an approach
to facilitate mapping of meanings across sources
and to extract data.

1. Introduction
The future scenarios for geographic information systems could be plausibly imagined as a potential single infrastructure working together in harmony with a wide variety of geographic-data. This would be highly beneficial to research scientists, decision-makers, disaster-management officers, forecast department officers and people. Today, a large amount of data of various geographic features is been collected by various departments. This data is often collected independently and hence it need not be of the same format. Moreover, the data that a user might request may not be available in a single data source and we would need to collect data from more than one data source. The information however retrieved from multiple sources would not be syntactically, structurally and semantically homogeneous and would be technically unusable. Manual integration would be a very tedious task and would be time-consuming especially at the time of disasters and emergencies.

Geographic information integration is the key to the aforesaid future scenario from the current state of heterogeneous data formats that prevail among several information systems. We can classify the heterogeneities that exist between different information systems as follows:

a) Data heterogeneity
It refers to heterogeneity among existing systems in terms of data representation and storage etc and can be further classified as below.
• Syntactic Heterogeneity - in which the information systems use different storage and representation techniques for data
• Schematic Heterogeneity - in which the same object in the real world is represented using different concepts and structures
• Semantic Heterogeneity - in which a fact can have more than one description or interpretation.
There is no consistency in the meanings

b) Functional Heterogeneity
Functional heterogeneity refers to the heterogeneity among existing systems in terms of service interfaces.

Open GIS Consortium’s (OGC) contribution towards integration by defining common specifications could not provide a seamless interoperability as it could resolve heterogeneity till the syntactic and functional levels. Seamless interoperability cannot be achieved by standardisation alone. It needs a tool and a methodology to provide interoperability at both schematic and semantic level.

This report describes efforts undertaken to develop a methodology and a tool as part of a research project at IIT Bombay with a sponsorship from Department of Science & Technology, Government of India, to resolve semantic heterogeneity in geospatial databases using ontologies. Experiments based on lexical similarity based ontology matching have been conducted. We describe the Framework for Interoperability of Geospatial Information using Ontologies (FIGO) developed as a tool. The workflow of creating and matching ontologies using FIGO is also described.

FIGO is a framework and also a tool. It allows definition of ontologies pertaining to various sources. It also allows users to define their ‘views’ of data based on their data needs, which may need extraction of data from one or more sources. A view is also treated like an ontology. Based on descriptions associated with the elements and using Wordnet (a web-based resource defining semantic relationships between English language words) to interpret these natural language descriptions, FIGO establishes mappings across ontologies. It associates a level of confidence among these mappings, which need to be reviewed by the user. Once the mappings are validated and accepted, FIGO can automatically generate ‘Wrappers’ which when applied to data from a source will transform that data into the form acceptable to another source or a user view.

FIGO is a web-enabled tool that allows registered users to create their own conceptualisation of data with the aid of the user-interface provided and stores them in the form of ontologies in FIGO. It also provides users to import their XML Schema in the tool and store it as ontologies in FIGO. Users can view, update and delete the sources at different levels of ontologies. FIGO tool also allows users to browse through ontologies of registered users that are available in FIGO. Fig. 1 shows a screenshot of FIGO for defining/browsing an ontology.
We explain each of these features of FIGO below.

2. Ontology Creation
Ontologies are used for making clear and explicit statements about entities in a particular domain.

An ontology can be defined as a formal, explicit specification of a shared conceptualisation.

Conceptualisation refers to an abstract model of how people think about things in the world, usually restricted to a particular subject area. An explicit specification means that the concepts and relationships in the abstract model are given explicit names and definitions. Definitions associate the names of entities in the universe of discourse with human-readable text describing what the names mean and says how a term necessarily relates to other terms.

For geographical domain, an ontology will define features, their attributes, and interrelationships in an application independent manner. An ontology will be like a conceptual data model for an application. In OGC approach, the ontology definition will lead us to the application schemas.

An ontology in our approach is defined in terms of classes, their properties and their interrelationships. We also allow Layer as a way of clubbing together a set of classes from application point of view (such as classes defining the transportation Layer). Classes in FIGO corresponds to geographic features such as Road, Bridges, River, etc. Layers may have sub-layers and classes may have sub-classes and attributes. Fig. 2 below shows elements of ontology.