Researchers' “Startup Readiness” in the Biopharmaceutical Domain Assessed Using Logistic Regression for Features of Their Papers, Patents, Institutes, and Nations

Abstract

This paper presents a method using logistic regression to predict and detect "startup readiness" of researchers in the biopharmaceutical domain, and to suggest determinants to improve their "startup readiness," using databases of start-up finances, research papers, patents, academic organizations, and national socioeconomics. This method sorts specific industry segments by which financing activities are active, and by which related growing research topics attract increased academic attention. In research domains such as the biopharmaceutical field, which include pursuit of fundamental scientific understanding and applications intended for immediate use, abundant startups with intense scientific linkage have attracted venture capital financing and entrepreneurship for further R&D opportunities and commercialization. We hypothesized that variables composed of several features of papers, patents, research institutes, and nations related to this domain can well reflect researchers' "startup readiness." Our logistic regression model based on our selected and constructed explanatory variables yielded good predictive and classifying performance, with an AUC value of 0.73. Results carried specific implications about what variables and their combinations demand attention, to encourage the "startup readiness" of researchers. More than conventional research methods, our computational approach might provide global, comprehensive, but convenient and real-time understanding of the "start-up readiness" of researchers in user-inspired fundamental research.