| Cote: |
1134 |
| Auteur: |
SAYOUR Fabio |
| Année: |
Juin 2019 |
| Titre: |
Cartography and Textual analysis : Similarity analysis adapted to the earth observation data |
| Sous la direction de: |
Dr Christian Kaiser |
| Type: |
Mémoire de master en géographie |
| Pages: |
46 |
| Complément: |
1 page d'annexe (tableau de données) |
| Fichier PDF: |
 Mémoire [1.5 Mo]
|
| Mots-clés: |
Chi2 / data-mining / similarity |
| Résumé: |
The present work describes the building of a geoportal based on the data from the forty projects belonging to the Group on Earth Observations (GEO). The web application seeks to meet two objectives. The first one is to expose at a glance the information related to the projects and their corresponding localisation over the world. This has let the people working at GEO to have a global view about the ongoing activities linked to the organisation and strengthen the link between the team working on those projects.
Following the idea of communication, the second goal was to build an algorithm that could analyse the similarities between each project based on their textual paragraph description found on GEO’s website. All paragraphs have been collected into a single textual corpus to build the contingency matrix and the Chi-2 distance calculation method has been applied on this matrix with the addition of a constraint parameter overweighting the contribution of rare terms within each paragraph. Hence, the algorithm has extracted for each project its most similar counterparts. Finally, this option of similarity retrieval has been added to the web application letting the user have for each project the choice to navigate to the three most similar projects by clicking on the corresponding buttons.
|