TY - GEN
T1 - Predicting daily consumer price index using support vector regression method
AU - Budiastuti, Intan Ari
AU - Nugroho, Supeno Mardi Susiki
AU - Hariadi, Mochamad
N1 - Publisher Copyright:
© 2017 IEEE.
PY - 2017/12/5
Y1 - 2017/12/5
N2 - Inflation rate could describe economic growth and it is usually used by policy-maker to determine a monetary policy. The Consumer Price Index (CPI) is one of indicator used to measure inflation rate. Until now, the inflation calculations and CPI prediction are conducted on monthly even though it is now likely to predict them on daily basis by utilizing online commodity price movement. Daily predictions could become a tool to analyze the real value of the market and will allow policy-makers to make better policy. This is a preliminary research to develop daily CPI prediction model by using Big Data. This paper discussed daily prediction model by using real-time data (daily commodity price and exchange rate) and SVR method. Build a model focused on accuracy and execution time. Grid Search and Random Search method were applied to select the best parameter for SVR model. In addition, we compared SVR method with linear regression and Kernel Ridge Regression method. The results show that the prediction model using SVR-kernel RBF has MSE value, 0.3454, less than other methods. Execute time for process data show that Kernel Ridge method has training time 0.0698s, little faster than SVR method 0.134s.
AB - Inflation rate could describe economic growth and it is usually used by policy-maker to determine a monetary policy. The Consumer Price Index (CPI) is one of indicator used to measure inflation rate. Until now, the inflation calculations and CPI prediction are conducted on monthly even though it is now likely to predict them on daily basis by utilizing online commodity price movement. Daily predictions could become a tool to analyze the real value of the market and will allow policy-makers to make better policy. This is a preliminary research to develop daily CPI prediction model by using Big Data. This paper discussed daily prediction model by using real-time data (daily commodity price and exchange rate) and SVR method. Build a model focused on accuracy and execution time. Grid Search and Random Search method were applied to select the best parameter for SVR model. In addition, we compared SVR method with linear regression and Kernel Ridge Regression method. The results show that the prediction model using SVR-kernel RBF has MSE value, 0.3454, less than other methods. Execute time for process data show that Kernel Ridge method has training time 0.0698s, little faster than SVR method 0.134s.
KW - Big Data
KW - Consumer Price Index
KW - Kernel Ridge Regression
KW - Linear Regression
KW - Support Vector Regression
UR - http://www.scopus.com/inward/record.url?scp=85045990406&partnerID=8YFLogxK
U2 - 10.1109/QIR.2017.8168445
DO - 10.1109/QIR.2017.8168445
M3 - Conference contribution
AN - SCOPUS:85045990406
T3 - QiR 2017 - 2017 15th International Conference on Quality in Research (QiR): International Symposium on Electrical and Computer Engineering
SP - 23
EP - 28
BT - QiR 2017 - 2017 15th International Conference on Quality in Research (QiR)
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 15th International Conference on Quality in Research: International Symposium on Electrical and Computer Engineering, QiR 2017
Y2 - 24 July 2017 through 27 July 2017
ER -
