A COMPUTERISED APPROACH TO COMPARISON BETWEEN AVERAGE PERCENTAGE AND RATIO-TO-MOVING AVERAGE METHOD FOR COMPUTING SEASONAL INDEX

Abstract

Although there exist other methods for estimating the Seasonal component , emphasis of this work is on Ratio-to-Moving Average method and the Average Percentage method. Attempt is made to compare the two methods with a view to establishing the supremacy of one method over the other. To effect the comparison, three linear equations (each of which is used to generate monthly trend values for a 10- yr period y ieldin g 120 data points) and three sets of Seasonal Index are used . Each o f the three equations is combined with each of the Seasonal Index t o yield nine different theoritical time series (T*S) . Each of the two methods under consideration is then applied on each of the nine theori tical series. Results (i . e estimated index values) obtained from each method is then compared (for each series) wi th the corresponding true index from which series in quetion was generated . The Mean- Squared Error (M . S . E) is c omputed f o r each o f t h e series under each of the two methods . The comparison is based on the M.S . E.. It is mathematically wise to state that the method that consistently produces minimal error is better.