executive summary
...e of consecutive data). By using this method the underlying trends or patterns otherwise hidden in the data maybe revealed. In the forecast for the number of registrations per week, we perform a 3-week and a 4-week moving average. The Exponential Smoothing method uses modified concepts of persistence of the observations and of error reduction and correction i.e., the forecast for next time ahead is the forecast at the current time corrected by some specified amount ‘alpha’ of the current forecast error, where alpha is the smoothing parameter. Here we perform two exponential smoothing method, one with alpha = 0.4 and another with alpha = 0.8. In both the cases, the exponential smoothing is initialized with a starting forecast set to the first week’s value of 22 registrations. Regression is the statistical method that fits a single straight line through all of the historical data using the method of least squares. The regression equation for our forecast is: Y = 2.103x + 19.133 Where y is the number of registrations calculated for week x. according to the regression model. The registrations are increasing at the average rate of 2.103 registrations per week. The coefficient of determination (R2) is 0.8523. Therefore 85.23% of the increase in the number of registrations from one week to the next is due to the passage of time according to the calculated regression equation. Almost all the increase in demand week after week is accounted for by regression according to the above model. The remaining 14.77 are not modeled by the regression due to possible random fluctuations. The correlation coefficient, which is .9232, indicates a nearly perfect positive linear association between registrations and weeks. By either measure, the regression is a very good fit to the given data. An accurate forecast is a forecast that agrees with the actual observation when the forecast time finally occurs in the future, the closer the agreement between the present forecast and the future actual observation; the more accurate is the forecast. This accuracy is measured by the overall accumulation of forecast error between the forecast value and the actual observation over all the forecast time periods. The accuracy method used here is the mean absolute deviation (MAD), which measures the average si...