Error classification

  • Sandrine PERALDI (ISIT)
    From raw automatic translation to professional post-editing: the case of financial translating
    2016, Vol. XXI-1, pp. 67-90

    This paper presents the implementation of an Applied Research Project aimed at evaluating the efficiency of a combined approach of machine translation and computer-assisted translation (CAT) toolsin the financial field. The purpose of the analysis, commissioned by a translation company specialising in regulated financial information, was to determine whether professional post-editing could offer a credible alternative to human translation from a qualitative and economical perspective, while helping the company streamline its outsourcing process. After reviewing the state-of-the-art in MT evaluation, we describe our research methodology (based on a human evaluation approach through the creation of a multi-faceted error typology) and the results of the study in the light of the company’s professional needs.