Inference versus evidence in reasoning remote sensing recognition, with an information foraging perspective

Authors:
M. Shoshany;Y. Cohen
Affiliations:
Department of Transportation and Geoinformation Engineering, Faculty of Civil & Environmental Engineering, Technion - Israel Institute of Technology, Haifa 32000, Israel;Institute of Agricultural Engineering, Agricultural Research Organization, The Volcani Center, Bet Dagan 50250, Israel
Venue:
International Journal of Remote Sensing
Year:
2007

Citing 8
Cited 0

A method for managing evidential reasoning in a hierarchical hypothesis space

Artificial Intelligence
Why expert systems fail

AI Expert
Expert system quality control

Information Processing and Management: an International Journal
Expert Systems: Design and Development

Expert Systems: Design and Development
Optimisation of multisource data analysis: an example using evidential reasoning for GIS data classification

Computers & Geosciences - Intelligent methods for processing geodata
Ontologies for geographic information processing

Computers & Geosciences - Intelligent methods for processing geodata
A national knowledge-based crop recognition in mediterranean environment

International Journal of Applied Earth Observation and Geoinformation
Geographic Data Mining and Knowledge Discovery

Geographic Data Mining and Knowledge Discovery

Quantified Score

Hi-index	0.00

Visualization

Abstract

Spatial, temporal and spectral complexity of remote sensing recognition tasks necessitates the use of Knowledge-Based Expert Systems (KBS). These systems are composed mainly of evidence and inference mechanisms: either domain-dependent inference (DDI) or domain-independent inference (DII). Selection of recognition strategies are typical of information foraging tasks and involve decisions regarding combinations of evidence and inference. This is highly dependent on the expected information gain (e.g. recognition accuracy and reliability) versus the cost/effort of constructing the evidential basis and the inference mechanism. This paper assessed a rule-based system (DDI) utilizing a sequent-oriented inference and a DII system utilizing the Dempster-Shafer evidential reasoning method. Quantification of evidence-inference-complexity-effort-accuracy relationships for a case study of land-use mapping on a wide regional scale allow a preliminary assessment of the relative performance of each strategy. Initial results indicate that a DII-based recognition system may function significantly better than a DDI-based system in large areas representing cases that had not been learnt during the evidence-extraction phase.