A combination projection — causal approach for short range forecasts†

† The original version of this paper appeared as ‘ A Combined Projection-Causal Approach for Short-Range Forecasts ’, Working Paper No. 527, Krannert Graduate School of Management, Purdue University. (September 1975)

Abstract

This paper illustrates a case study of forecasting daily traffic levels at branch banks, where many behavioural and business factors are present. Many influences are not equally spaced over, which reduces the effectiveness of traditional time series approaches. To handle this problem, a univariate time series ARIMA model is developed mid then dummy variables are added to incorporate exogenous effects that are not captured by the projection ARIMA model. The results indicate that a more adequate representation of the customer traffic pattern has been obtained by combining the two modelling approaches

Notes

† The original version of this paper appeared as ‘ A Combined Projection-Causal Approach for Short-Range Forecasts ’, Working Paper No. 527, Krannert Graduate School of Management, Purdue University. (September 1975)