Abstract: The traditional asset pricing model prices assets from the perspective of the market, but asset prices cannot be fully explained because the investor interaction behavior behind the market is closely related to the stock asset price. This paper uses the complex network method to model the investor interaction relationship, describes the interactive patterns of online investors through the motif structure of the investor interaction network in the stock forum. Based on the Fama-French three-factor model, the investor interaction factor is added to construct a four-factor pricing model. It is found that investor interaction factor has certain pricing power, and the pricing effect of the four-factor model with interaction factor is stronger than that of the traditional three-factor model and the four-factor model with momentum factor. This study provides empirical evidence for the market value of unstructured text information under the big data framework.

Key words: investor interaction, asset pricing, complex network, network motif, factor investing