Venture Capital Investment
Traditionally, venture investors (e.g., business angels, venture capitalists, private equity investors) make investment decisions based on their past investment experiences, social connections and/or qualitative assessment on startups. Current studies on venture capital investment, from finance and/or managerial perspectives, are mostly based on post hoc methodologies (e.g., interviews and surveys). However, people’s retrospection is subject to rationalization and recall biases. Therefore, the entrepreneurial finance industry has an active call for quantitative and methodologically sound studies on venture capital investments. To this end, our research aimed to apply cutting-edge data analytics technology to help venture capitalists make better and smarter investment decisions. At first, we attempted to address the problem from a personalized portfolio perspective. We developed a Probabilistic Latent Factor model to estimate investors’ investment preferences collaboratively. By taking account of potential investment returns and risks, we utilized Modern Portfolio Theory to optimize the startup portfolio generated from the investment preference model. As a result, this strategy can maximize investment returns along with the potential risks suppressed, and in the meantime, meet the investment preferences of the investors. Following this work, we also attempted to study the impact of prominent social ties between members of VC firms and start-up companies to the investment decision-making process, which was critical but attracted little attention. Specifically, we developed a Social-Adjusted Probabilistic Matrix Factorization (PMF) model to exploit members’ social connections information from VC firms and startups for investment recommendations. Our model brought in more flexibility, and the results inherently provided meaningful managerial implications for the controllers of VC firms and startups.
上午 10:00 ~ 11:30
Hao Zhong, The State University of New Jersey