基于POT方法的极值理论在基金净值预测中的应用

Application of extreme value theroy in predicting fund based on POT method

为对基金净值数据进行建模,根据基金净值样本数据的尾部特点,建立极大,极小值分布的GPD模型,运用POT方法确定临界值,进而对参数进行估计,并对模型进行检验.最后,运用建立的模型对一些极值点进行预测.所得结果很好地描述了数据特点,对极值点的预测符合实际.

For moding the data of net value of fund, according the tail feature of sample data of net value of fund, the paper estaiblishes the GPD model of maximum and minimum distrbution, using the POT method to determine the critical value, then estimates the parameters, and the models were tested. Finally, using the model to predict extreme points. The results well describes the feature of data, the forecast of the extreme points accordes the reality.