DocumentCode :
552975
Title :
Which is better for inflation forecasting? Neural networks or data mining
Author :
Somaratna, P.E. ; Arunatilaka, S. ; Premarathna, L.
Author_Institution :
Sch. of Comput., Univ. of Colombo, Colombo, Sri Lanka
fYear :
2010
fDate :
28-30 June 2010
Firstpage :
116
Lastpage :
121
Abstract :
In present context, Information Technology (IT) is used in every field whether it is Business, Social Service or Entertainment. However, application of IT in economic field is very limited. Stability and healthiness of a country´s economy directly connects with the accuracy of forecasted Inflation rate. With the recession raised at end of year 2008, world communities pay much attention on inflation and put huge efforts to predict it accurately. Neural networking and data mining are two IT techniques that are commonly used in forecasting and hidden pattern recognition. We have applied both of these techniques to the theoretically sound data set with the intention of identifying most appropriate IT forecasting technique for forecasting the inflation rate. Since forecasted inflation rate directly link with country´s monetary policy, accuracy of predictions is very significant. Further continuity of government as well as the economy depends on these decisions. Through this study we were able to identify appropriate characteristics of neural networks and Data mining techniques in case of forecasting inflation rate with high accuracy.
Keywords :
data mining; economic cycles; economic forecasting; government policies; inflation (monetary); macroeconomics; neural nets; pattern recognition; IT forecasting technique; country monetary policy; data mining; economic field; hidden pattern recognition; inflation forecasting; information technology; neural networks; recession; Biological neural networks; Data mining; Data models; Forecasting; Mathematical model; Predictive models; Training;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Information Society (i-Society), 2010 International Conference on
Conference_Location :
London
Print_ISBN :
978-1-4577-1823-6
Electronic_ISBN :
978-0-9564263-3-8
Type :
conf
Filename :
6018807
Link To Document :
بازگشت