Discovery of chorems by spatial data minig

Ibtissem Cherni 1, 2
1 BD - Base de Données
LIRIS - Laboratoire d'InfoRmatique en Image et Systèmes d'information
Abstract : Our dissertation motivation is basically based on the "chorem" concept. This neologism, invented by Roger Brunet, points out to both an element and a structure of the system. A "chorem" starts from a simple geometrical element, such as a dot, a line, a vector, a network.. to form much more complex semantics. These combinations create all types of interpretation, starting from a simple representation of an important place through a given point, to flows of exchanges which exist between zones, with the help of circles, lines, colours, textures, arrows, etc. The aim of our work is to define a cartographic solutions in order to better represent geographic information extracted from the content of geographic database which refers to statistic objects and dynamic phenomena at the same time. The visual representation, in a simplified map, of the information extracted from this analysis, becomes a solution to an even more complex problem, especially when the domains at stakes are political, economic, and demographic. We propose a solution based upon the concept of the "chorem" and its capacity to epitomize the scenarios involving statistic objects and dynamic phenomena, relating them to visual schematic rating. The objective of our methodology is to extract the patterns which may serve as a basis to onstruct visual display of a geographic database. These patterns are the following: clusters, facts, flows, co-localizations, topological constraints, and exterior information. Nevertheless, the number of patterns extracted during the first phase is seldom important, so, we procede to reduce it being based on the elimination of the unnecessary knowledge from the view point of an expert, at the same time, we exclude those which are redundant. As far as the phase of visualization is concerned, we propose two alternatives: results based on the technique of the concepts lattices and the direct visualization as chorems. Generally, our approach comprises 3 phases: 1- The first as being the extraction of patterns through spatial data mining 2- The second is dedicated the identification of the most important patterns 3- And the last phase concerns visualization of the results. To validate this approach, a prototype was implemented.
Document type :
Theses
Complete list of metadatas

Cited literature [76 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/tel-02000009
Contributor : Sylvie Servigne <>
Submitted on : Thursday, February 7, 2019 - 10:19:32 AM
Last modification on : Wednesday, November 20, 2019 - 3:14:12 AM
Long-term archiving on: Wednesday, May 8, 2019 - 12:11:16 PM

File

These-IbtissemCherni.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-02000009, version 1

Citation

Ibtissem Cherni. Discovery of chorems by spatial data minig. Base de données [cs.DB]. INSA Lyon; ISAMM, 2015. Français. ⟨tel-02000009⟩

Share

Metrics

Record views

41

Files downloads

88