摘要:人民币兑换美元汇率的波动、走势的分析与预测在金融研究中具有重要的研究意义。本文通过收集并分析2010年至2012年人民币兑换美元的季度数据、月度数据、周数据，通过自相关系数、偏自相关系数检验它们的Markov性，用马尔可夫链预测模型进行预测，结果发现：季汇率预测值，月汇率预测值，周汇率预测值随着样本数量的增大，汇率收盘价预测值波动越来越小；随着样本量的增大，达到稳定值的转移步数越来越大；随着时间的推移，人民币兑换美元的汇率收盘价将以很大的概率落入第一个状态区间。

关键词：Markov性检验；Markov预测模型；人民币兑换美元的汇率；转移矩阵

ABSTRACT:Analysis and prediction of the renminbi to us dollar exchange rate fluctuations and trends has important significance in financial research. In this paper, through the collection and analysis of 2010 to 2012, the yuan against the dollar, the quarterly data of monthly data, weekly data, and using the autocorrelation coefficient, partial correlation coefficient test Markov property of their. As predicted by Markov chain prediction model, The results found: Quaternary rate forecast, monthly rate forecast, week rate forecast with increase of the number of samples, exchange closing price prediction value fluctuation smaller. With the increase of the number of samples, to transfer step stable value increasingly large numbers. With the passage of time, the renminbi to us dollar exchange rate closing price will fall into the first state interval with a great probability.

Key words: Markov property test; Markov prediction model; RMB exchange rate; Transition matrix