Development of a data-driven scientific methodology: From articles to chemometric data products

Information and data science algorithms were combined to predict the outcome of an experiment in chemical engineering. Using Scientific Method workflow, we started journey with formulation a specific question. At research stage, common process querying reading articles on scientific databases was substituted by systematic review built-in recursive mining method. This procedure identifies community knowledge key concepts experiments that are necessary address formulated A small subset relevant from very topic among thousands papers identified while assuring loss least amount information through process. The secondary dataset bigger than individual study. revealed main ideas currently under study optimal synthesis conditions produce substance. Once step finished, experimental compiled prepared for meta-analysis using supervised learning algorithm. is hypothesis generation stage whereby transformed into about particular reaction. Finally, predicted sets desired compound validated laboratory.