关键词: 流感预测, 时间序列, 变量选择

Abstract: In the era of big data, people argue that the traditional statistic has lost its superiority in data analysis. In this paper, we take ARGO model as an example to introduce the applications and achievements of statistical methods in big data analysis, and put forward the potential improvements from statistics’ point of view. The analysis of ARGO model shows that many intrinsic problems in big data can be resolved by statistical methods and the statistical law contained in the data is still the greatest value of data mining, which means that the statistical thinking can only become more and more important in big data analysis. However, in the face of big data with complex structure and diversity of sources, statistical methods also need further exploration and try to seize the new opportunities for development and rise to new challenges.

Key words: Prediction of Flue Trend, Time Series, Variable Selection