New weighting system for e-readiness indicators based on indicator importance measurement by the random forest algorithm

Weighting the indicators was always a difficult step in building the composite indicators. In the e-readiness assessment approaches, where there are several indicators of different categories, the weighting methods used are not effective enough to assess the importance and the real priority of the indicators. In this article, we have presented two contributions: the first consists in combining subjective weighting with objective weighting to build a complete and optimal weighting system. the second contribution aims to propose a new statistical method based on the random forest algorithm to measure the importance of indicators and calculate objective weighting. A case study on the Internet Inclusive Index of 2019 is illustrated to assess the effect of the new weighting system on the scores and ranking of 100 countries. © 2005 - ongoing JATIT & LLS