Dataset for "Balancing factoid retrieval and candidate recommendation to identify toponym origins in gazetteers"

Résumé fourni par la source

Data and code to reproduce the figures and results in the paper "Balancing factoid retrieval and candidate recommendation to identify toponym origins in gazetteers" for the International Journal of Geographical Information Science. The instructions and descriptions are provided in the README file.Unzip the data in a folder where your have the full rights of reading/writing (e.g. ~/Documents) otherwise some files might face conflicts (Python, QGIS file).Optional: If you want to go further by regenerating all the data from scratch, you must also unzip the compressed file "raw_code_models.zip". Follow the instructions in the README in the extracted sub-folders. However, the used libraries might be experimental: we highly recommend to work with a Unix system and to follow the instructions of each library (if debugging is needed).The data in "CODE_Spatial_RAG/data/EXP2/RAG_02_after_rewriter/kg_view_geo.zip" (>50k files) is only useful to extract if you want (a) to analyze the data after the re-writer or if you want to (b) copy-paste the results when regenerating the data from scratch. Some paths were automatically created and are not compatible with a Windows system (long / invalid characters) but only with a Unix system. You can also "skip" the unsuccessfully uncompressed files.