Sophistical procedure are often required for collaborative geospatial problem-solving and decision making

while most researchers in GIS-relevant studies pay much attention to the generation of geographic datasets, information on how these data are derived is often ignored.

spatial privenance defined here as lineage (世系) and workflow information of spatial data manipulation and related analysis.

• to estalish a scalable and extensible architecture of provenance-aware
• to develop a provenance model suited to handle information of spatial data derivation
• to provide services to facilitate the storage and query of spatial provenance repository for GIS applicaiton, developers, and users.
  

Provenace might be used to determine lineage, by tracking steps involved in sourcing, moving, and processing of data or how it arrived in a particular database.

By recording provenance it is possible to achieve a sound lineage record that ensure reproducibility of meaningful data transformation.
 
Three major approaches have been developed for provenance representation

• annotation

             record provenance information at defferent granularity level, the provenance granularity depends on application domain requirements, and the cost of collection, storage and processing

• inversion

             function defined along with output data that are sufficient to identify data sources, and allow a compact representation of provenance that enables scalability.

• virtual data
  

             a catalogue is maintained for representing data derivation procedures and derived data, providing sufficient information to re-compute output as needed.

To record and retrieve provenance from the semantic and spatial repository necessitates the development of a consistent machanism to collect and store provenance.

The OPM defines agent, artifacts and process as basic entities and uses a directed graph to capture the casual dependencies between the entities forming relationships.

In the OPM, casual dependencies between process and data from a provemence graph. such graph can be used to represent workflows as directed acylic graphs (DAG) that express causal relationships between the preocesses with the datasets acting as links.

The causal relationships associated with data derivation is validated by using start-time and end-time recorded for each process. The OPM highlights the utility of timestamps in understanding causality.

source, units, formats, temporal reange, and alternat data location associated it.

Tupelo is used as the semantic content repository for storing, retrieving and accessing provenance. Tupelo is middleware that provides a Web access protocol and JAVA API that interface with and RDF mapping of the Open Provenance Model.

Provenance data is stored using semantic Web technologies based on the RDF, and is backed with standard storage technologies (eg.database) and RDF stores (e.g. Sesame)

Provenance in GIS can represent causal realtionships among spatial ddata and associated analysis procedures.

Oppertunities
     using provenance to enable and facilitate cutting-edge GIS capabilities(e.g. data-intensive and complex spatial analysis in distributed and collaborative GIS)
 and to improve the effeciency and quality of spatial analysis for coordinate development of geospatial knoledge.

Challenges:

1) developing a comprehensive GIS ontology, in particular, space-time ontology to enhance the geospatial capabilites of either provenace-aware GIS or generic provenance systems
2) adapatively and intelligently controlling proveance granularity and recording based on different application requiredments
3)gaining better understanding about cost-benefit tradeoffs between tight- and loose-coupling stratigeis for integrating provenance capalitities within GIS









转载本文请联系原作者获取授权，同时请注明本文来自郭旦怀科学网博客。
链接地址：https://wap.sciencenet.cn/blog-204718-277283.html

上一篇：Reasearch Spirit
下一篇：RDF