While the potential data landscape in econometric research is undergoing dramatic changes, terminologies such as “Big Data” and the associated discipline of “Machine Learning” (ML) have so far received little attention among econometricians. In this chapter, we illustrate the potentials for and promises of quantitative entrepreneurship research to benefit from the availability of unprecedentedly rich datasets and non-traditional data sources such as text, video, or image data. However, we also highlight that such datasets are in need of new approaches, both methodological and epistemological. We proceed by introducing the ML approach to quantitative analysis geared towards optimizing predictive performance, and contrast it with standard practices in econometrics that focus on producing good parameter estimates. The chapter also introduces ML techniques such as out-of-sample model validation and variable selection, as well as regularization procedures. We further illustrate and exemplify these principles by exploring popular ML techniques such as classification and regression trees, artificial neural networks, and vector space models for natural language processing. We provide guidance on how to apply these techniques to quantitative research in entrepreneurship and point towards promising avenues of future research that could be enabled by the use of new data sources and estimation techniques.
|Titel||Handbook of Quantitative Research Methods in Entrepreneurship|
|Redaktører||George Saridakis, Marc Cowling|
|Forlag||Edward Elgar Publishing|
|Publikationsdato||26 jun. 2020|
|Status||Udgivet - 26 jun. 2020|