For the problem that the commonly used risk evaluation model only uses cross-sectional data and the trend
information of evolution is not fully reflected in the risk evaluation models, a risk evaluation model based on the risk of
specific indexes is proposed. For the state variables that can be reflected with cross sectional data, the risks are measured
through Sigmoid function. For the process variables that need to be reflected with time series data, the risks are measured by
using two-attribute models and value-at-risk models. By considering the mean, variance and skewness, the time series data
are mapped to a cross section data such that the risk evaluation model has the ability to handle the dynamic information. A
specific example shows the feasibility and effectiveness of the proposed method.