Measurement and Analysis of Industry Risk-Return Value Measurement and Analysis in the Context of Big Data Technology - A Study Based on VaR and VaB

Based on VaR and VaB research methods, this paper constructs a regression model based on industry risk-return value metrics using a Bayesian-expected quantile regression algorithm in the context of big data. The parameters of the Bayesian model are estimated using the decision function, and then the metric effects of the model are evaluated using AIC and BIC. Finally, the relationship between industry risk and return is evaluated using the Bayesian-Expectation Quartile regression model. Under the Bayesian-Expectile model, the difference between the correlation coefficients of AIC (−0.7431, −0.9563, −0.7872, −0.9137) and BIC (−0.0547, −0.0768, −0.0559) is relatively small, reflecting the fact that the Internet financial innovations have not significantly pushed up the risk-free rate of return of the society while pushing up the society’s financing cost. In addition, the three models used in this paper have close EvaR prediction results for information technology, daily consumption and telecommunication services, indicating that the credibility of the industry risk prediction results proposed in this paper is high. The relationship between idiosyncratic volatility and the adjusted return for the next period is related to the time-varying nature of the characteristics of individual stocks, as well as the individual stocks of industry-risk firms. Incorporating factors that affect conditional equity premiums into the time-varying conditions of individual stocks can more accurately reflect the risk-taking of individual stocks of risky firms.