At the Applicant Webinar of August 9 2012, it was mentioned that grouping similar answers is a step involved in the revision of gTLD applications. This website shows a possible answer similarity grouping. By using different similarity criteria, it is possible to group questionnaire answers in ways that provide useful insights. If you would like more information or a different similarity rendering of the questionnaires do not hesitate to contact me. I will be implementing additional features as time and resources allow.

Arnoldo J. Muller-Molina

Brief Method Description

The BM25 technique is widely used in information retrieval systems. It has been employed by Murata et al. in patent analysis tasks. A minimal modification must be performed in order to use the BM25 on a similarity index. BM25 Parameters were set to b=0.5 and k1=1.2. Please note that the groups created have overlaps, that means an answer may appear in more than one group of answers.

About this Project

This analysis has been created with the R-01 similarity engine developed by my start-up simMachines. The project has been partially sponsored by the Costa Rican Ministry of Science and Technology (MICIT) and the National Council for Scientific and Technological Research (CONICIT) through the Propymes grant FP-034-12.


Address:971-3100 Santo Domingo Costa Rica
Telephone:+506 8542-1956