Accurate measure of the risk of stock information is of great significance for asset pricing，risk management，and the measure of market performance． To overcome the shortcoming of assuming constant probabilities of the states of news and symmetric order-flow shock of the model of Duarte and Young，we extend the methodology of Duarte and Young． We model the probabilities of the states of news and the probabilities of symmetric Order-Flow shock by using trade volumes． Our method allows both the probabilities of the states of news and the probabilities of symmetric Order-Flow shock to vary． Our APIN and PSOS estimates can be computed daily as well as over intraday intervals． Then，we select some actively traded stocks to do an empirical research and our model are compared with the model of Duarte and Young． Our empirical results show that our model provides a better fit for our data and has better explanatory power for spread． At last，we do an empirical research for the inter-day pattern of information risk．