Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
We propose a more flexible and useful model, BP-CVaR, to estimate conditional value-at-risk (CVaR) using back propagation (BP) algorithm, which can capture the change of markets and use the information to adjust the next CVaR result. We use three samples including S&P 500 index, Nasdaq index, DIJA index and compare the results by back-testing approaches. We find that (i) BP-CVaR is more reliable and has higher accuracy than Monte Carlo CVaR model and (ii) BP-CVaR can react to the change of markets more quickly than traditional CVaR model.