分位数回归在时间序列中的应用

The Application of Quantile Regression in Time Series

文章利用分位数回归和时间序列相结合的方法对澳大利亚月度红酒销售量数据进行建模和预测,得出的模型能很好地描述出月份对于红酒销量变化范围的影响.当自变量时间对因变量红酒销量的分布产生不同的影响时,相对于最小二乘回归系数得到单一结果来说,利用分位数回归得到的时间序列模型能更好地利用数据里的信息,得到比较全面的预测结果.

Taking advantage of the application of quantile regression in time series,we predict the future red wine sales of Australia.Models getted can give a good discription of affection of month to sales.Comparing to single result from the Least-square method,quantile regression can make full use of the data and give more comprehensive results.