Analyzing and Providing Comprehensive Feedback for French CVS with Readerbench

Gabriel Gutu, Ionut Cristian Paraschiv, Mihai Dascălu, Gabriel Cristian, Stefan Trausan-Matu, Olivier Lepoivre

    Research output: Contribution to journalArticleAcademicpeer-review

    11 Downloads (Pure)


    In their everyday activities, recruiters are faced with the difficult task of analyzing and judging the quality of a wide range of CVs. Both the content quality and the visual hues, such as colors and their overall structure, need to be considered. This article enhances previous researches with a larger dataset, refined indices, and a more advanced technique of parsing the input documents. After applying various processing techniques from ReaderBench, an advanced Natural Language Processing framework, on a manually annotated dataset of 96 positive and negative French CVs, several writing indices were determined and filtered by leveraging statistical analyses. In addition, our experiment introduces a web application in which users can submit, gather an evaluation, and acquire valuable feedback on their CV.
    Original languageEnglish
    Pages (from-to)17-28
    JournalPolytechnical University of Bucharest. Scientific Bulletin. Series C: Electrical Engineering and Computer Science
    Issue number2
    Publication statusPublished - 2018



    • CV analysis
    • CV assessment
    • text cohesion
    • textual complexity
    • Natural Language Processing

    Cite this