LPPL模型算法的初步探讨

Study of Fitting Algorithm of LPPL Model

LPPL模型是对导致金融市场崩溃的金融泡沫进行诊断与预测的工具.介绍了LPPL模型并对模型的拟合算法进行了梳理,提出了基于网格的预搜索Nelder-Mead Simplex方法并进行了实证.结果表明,模型对恒生指数的泡沫与崩溃预测较为理想,对所有的泡沫崩溃过程均可得到在可接受的范围内的参数值.

LPPL model is an important tool in the diagnosis and prediction of financial market bubbles that precede large market falls or crashes. This paper introduces the LPPL model and its underlying mechanism, and puts forward a preliminary search procedure based on a grid to pro- vide seeds for the Nelder-Mead Simplex after an introduction of fitting algorithms. Empirical re- sults show that the model of the prediction of bubbles and crashes of HSI is ideal for all the bub- bles, and parameter values in the acceptable ranges are observed.