Reem Ismail


University / Organisation : Dauphine

Paper or project ? project

Title : A hybrid method to measure startup novelty

File : Download

Abstract : In this research work, we will study and develop new methods to measure and detect innovation and novelty based on data-science techniques and textual datasets. For many decades, innovation has been analyzed in a variety of disciplines, and various approaches have been developed with respect to its measurement, most notably including patent databases. More generally, these methodologies always have to deal with noisy and imperfect datasets and it would, in any case, be particularly useful to supplement them with additional and complementary approaches, especially whenever academics and practitioners are willing to discriminate among innovations in terms of type, impact, or else. In this context, the recent improvements in data science have precisely opened the way for the development of new methodologies to measure and detect innovation based on textual descriptions, based on NLP techniques. These methods can be developed on various datasets (corporate projects, publicly funded projects, policy projects, startups, etc.) and of course also in relation with metadata notably of a financial nature