Prediction of Financial Crime Based on the Improved Grey Model

Crimes forecasting commonly can be divided into two categories: one is historical data based on the criminal's own factors; second is correlative data according to the social factors affecting the crimes rate, such as degree of economic openness, the income gap, etc. However, in real crimes data, it is common that the data consists of the both. A single model may not be sufficient to identify all the characteristics of the data. In this paper, we firstly forecast financial crimes in Beijing-Tianjin-Hebei of China by a basic model using GM (1, 1), then adjust the errors which are caused by other social factors, finally we establish a revised forecasting data series by adding the errors to the raw forecasting, which in fact, is equivalent to adjust the forecast model of financial crimes twice.The experimental results shows that the improved model is able to produce greater accuracy as compared to the individual model be.