Dynamic financial contagion prediction model based on fuzzy information granularity SVM

摘要

Contagion time prediction is an important research topic in financial crises. This article put forward a prediction model of contagion time based on fuzzy information granularity SVM. It uses granularity fuzzy and SVM to estimate the bounds of stock index, and further forecast the similarity index. The predicted contagion time from the United States to the United Kingdom, Germany, Frence and China are tested, and compared with the real ones. The empirical analyses comfirm that the model is a feasible method to predict the financial contagion arrival time.