Abstract: This paper states the basic concept of deception and analyses the causes that give rise to the risk of deception. The methods of identification, analyzing and estimation of deception risks are discussed. The shortcomings of the traditional models for evaluating deception are also discussed. A method based on the rough neural network for analyzing deception risks is suggested and an example is given to validate the algorithm. The results show the feasibility and validity of the method. The research work furnishes a basis for further study of applying rough neural network to the analysis and prediction of deception risks.

Key words: risks, rough set, neural network, predicting, deception risks