ARPEGE: Robust Analyses for Geographical Explorations
Author of the model: Didier JOSSELIN
General themes of interest
- Exploratory Spatial Data Analysis (ESDA)
- Analysis of the morphology of the geographical objects
- Spatial forms, composite multiscale geographical objets
- Organization and topology of geographical objects
Goals of ARPEGE
- Analysis of the structural, functional, topological and statistical relations between geographical objects
- Critic and educated analysis of models
- Answering the question: is it possible to detect interacting elements in order to define understandable geographical objects for an expert?
Specific modules
=> Spatial autocorrelation unit: measure of local autocorrelation. Are close elements or objects rather different or similar?
- Available analyses: calculation of the Moran and Geary coefficients
=> Co-occurence computation unit: co-occurences analysis of a phenomenon, detection of spatial forms. How to measure and highlight the distribution of the elements or objects the ones compared to the others?
- Available analysis: omnidirectional co-occurences applied to the image processing
Theoretical references
ARPEGE is built on the basis of three paradigms:
- the systemic paradigm: highlighting of the relations between objects
- the object paradigm: properties of formalization and programming on these objects and their relations
- the "exploratory" paradigm: statistical approach, essentially graphic and without constraints
ARPEGE is developed in a statistical environment (L. Tierney, Lisp-Stat environment).
Inputs
Values, tables
Maps
Images
Outputs
Values, tables
Curves
Maps
Download
Comments
Only simplified examples of the modules of Space Autocorrelation and calculation of Matrices of Co-occurence are provided, in accordance with the contents of the presentation of the formation. The provided applications are built voluntarily with a didactic aim. As XLISPSTAT is a free software, all the applications developed in this environment can be diffused freely by the users. Some rules must however be complied with to support the authors in "the collective adventure of the free software": any use of XLISPSTAT requires to quote the authors (Betz, Tierney), idem for all the affiliated software: Arc (Cook Weisberg), VISTA (Young), LiveMap (Brunsdon), ARPEGE’ (Josselin), MIASME (Banos), etc. It is formally interdict to take again the code and to use it at any commercial purpose, no support nor maintenance private individuals are planned for the use of this software, all the remarks bound for the authors are the welcomes. Data-processing type of code: methods (in a strict sense of Directed Object) associating objects dynamically
References
L. ANSELIN, 1995. Local indicators of spatial association, LISA, Geographical Analysis 27, 93-115
R. M. HARALICK, 1979. Statistical and structural approach to textures, Proc. IEEE, vol. 67, 786-804
D. JOSSELIN, 1999. A search geographical composite objects with the prototype ARPEGE, Revue internationale de Géomatique, «Data mining spatial», Editeur K. Zeitouni, vol.9, n°4, 489-505
D. JOSSELIN, J. BOLOT, 2000. A semi-automatic method to build territorial partitions, Proceedings of the 5th International Conference on GeoComputation, University of Greenwich, United Kingdom, 23-25 August 2000 Click here to access
J.-L. LE MOIGNE, 1990. La théorie du système général, théorie de la modélisation, PUF
P.A. MULLER, 1999. Modélisation objet avec UML, Eyrolles
L. TIERNEY, 1990. Lisp-Stat, an object oriented environment for statistical computing and dynamic graphics, Wiley-Interscience Publication, John Wiley and Sons, NewYork
J.W. TUKEY, 1977. Exploratory data Analysis, Addison-Wesley
Key word(s) linked to this articleSoftware » download
This article last updated Sunday 17 September 2006. by Josselin Didier