Building CPI Forecasting Model by Combining the Smooth Transition Regression Model and Mining Association Rules

  • Do Van Thanh
  • Cu Thu Thuy
  • Pham Thi Thu Trang

Abstract

Inflation forecast plays a very important role for  stabilizing  the  economy.  In  Vietnam,  inflation  is measured  via  consumer  price  index  (CPI).   CPI’s changes  depend  on  many  factors  in  which  the merchandises’  price  changes  are  direct  factors  and those changes are not difficult to observe. The aim of our research is to propose a CPI forecasting model   based  on  the  change  of  merchandise  pricing since  such  a  model  has  not  been  built  so  far.  A comprehensive  study  has  been  carried  out  to understand the effects of price changes of merchandises on  CPI.  After  that  Nonlinear  Smooth  Transition Regression  Model  and  Mining  Association  Rules  are applied to build the model. The model parameters were configured  and  justified  using  actual  data  collected  in two years 2008-2009. The results showed the accuracy of the  model  for  CPI  forecast  in  Vietnam  and  the  model can  also  be  used  to  predict  the  price  changes  of merchandises.
Published
2010-10-28
Section
Regular Articles