ABSTRACT
Venture capital (VC) is financial capital provided to early-stage, high-potential growth start-up firms. VCs are unsung heroes behind high-tech firms, such as Google, PayPal, and Alibaba, especially when the firms are in their infancy. This chapter considers the problem of coinvestment prediction in the VC network and perform a series of observations of the data. It introduces the dataset and describes feature design and feature selection by group Lasso. The chapter presents observation of the prominent features selected by group Lasso and proposes structural balance–based factor graph model and learning algorithm. It also presents experiment results and detailed analysis and explores another dataset that focuses on the start-ups in China. The chapter provides an illustration of investment and coinvestment, present the formal definition of coinvestment, and then propose a formal description of the problem. It outlines the problem in the context of VC to keep things concrete.
