A MIDAS modelling framework for Chinese inflation index forecast incorporating Google search data

作者：

Highlights：

• We propose a MIDAS framework which incorporates Google search data to predict Chinese inflation index.

• We extract five groups of keywords, which include both subject terms and trends terms to construct search data.

• The high frequency part of Google search data is directly used in MIDAS model to predict monthly CPI.

• The MIDAS model outperforms other benchmark models in CPI forecast.

摘要

•We propose a MIDAS framework which incorporates Google search data to predict Chinese inflation index.•We extract five groups of keywords, which include both subject terms and trends terms to construct search data.•The high frequency part of Google search data is directly used in MIDAS model to predict monthly CPI.•The MIDAS model outperforms other benchmark models in CPI forecast.

论文关键词：Inflation index forecast,Consumer price index,MIDAS modelling framework,User generated content,Google search data

论文评审过程：Received 30 September 2013, Revised 31 December 2014, Accepted 2 January 2015, Available online 9 January 2015.

论文官网地址：https://doi.org/10.1016/j.elerap.2015.01.001