Neuro fuzzy modelling for prediction of consumer price index

Thumbnail Image
Journal Title
Journal ISSN
Volume Title
University of Dar es Salaam
The existence of efficient economic forecasting model is crucial for financial policy makers in a territory. The economic indicators such as Consumer Price Index (CPI) have been used frequently in predicting future economic wealth of the country. Recently, central banks such as the Bank of Tanzania, have adopted an inflation targeting monetary policy regime, which accounts for high requirement for effective prediction model of consumer price index. There are number of research works conducted to predict CPI using various techniques, however, their prediction accuracy is still low, which raises the need for improvement. This dissertation presents an artificial intelligent technique that combines neural networks and fuzzy logic (neuro fuzzy) to predict a univariate time series CPI data. The simulation experiments were conducted on Matrix Laboratory (MATLAB) using monthly CPI data taken from Tanzania National Bureau of Statistics from January 2000 to December 2015. Ninety five percent (95%) of the data were used to train the model and five percent (5%) for testing. Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE) were used as error metrics for model evaluation. The results showed that, the neuro fuzzy model having an architecture of 5:74:1 with Gaussian membership functions (2, 2, 2, 2, 2), provides RMSE of 0.44886 and MAPE 0.23384, which is better when compared to existing research studies.
Available in print form, EAF collection, Dr. Wilbert Chagula Library (THS EAF QA 76.87.T34A42)
Neural network, Computer science, Tanzania
Ambukege, Godwin (2017) Neuro fuzzy modelling for prediction of consumer price index, Masters dissertation, University of Dar es Salaam, Dar es Salaam