Forecasting Techniques CMA Part 1 Question 3 1 : As part of a risk analysis, an auditor wishes to forecast the percentage growth in next months sales for a particular plant using the past 30 months sales results. Significant changes in the organization affecting sales volumes were made within the last 9 months. The most effective analysis technique to use would be a) Unweighted moving average. b) Exponential smoothing. c) Queuing theory. d) Linear regression analysis. (CIA Adapted) Other Measures Used in Regression Analysis Forecasting In addition to the coefficient of correlation, r, several other measures are used to evaluate the precision and reliability of a regression forecast. These would be performed either at the same time as or following the actual regression analysis: • The size of the standard error (SE) of the estimate must be interpreted in relationship to the average size of the dependent variable. Hence, the units of the error may seem large (i.e. 10,000 units), but in comparison to the size of the variable (100,000,000 units), that percentage of error would be small. If the standard error of the estimate is around 5-10% or less of the average size of the dependent variable, we can be confident that the regression analysis is fairly precise. SE aids in establishing a confidence interval around the forecasted value. This range of confidence at 95%, 97% or 99% tells the likelihood that the random variable x can be found within the defined interval. The higher the percentage of confidence (99%), the higher the assurance that the value will fall into the range. However, the precision of the forecast suffers because of the broader range of values that are in the range of possible results. • The coefficient of determination is the square of the coefficient of correlation. It is represented by the term R2, or r, and it is the percentage of the total amount of change in the dependent variable that can be explained by changes in the independent variable. R2 is expressed as a number between a and 1. In a regression with a high r, the data points will all lie close to the trend line. In a regression with a low r, the data pOints will be scattered above and below the trend line. An r above .50 would indicate that the forecast yielded by simple linear regression analysis should be meaningful. In the above example, r, the coefficient of determination, is .94786 trend line were downsloping and the coefficient of correlation were -.94786, for example, the coeffi cient of determination would still be .8984, since squaring eliminates the negative value. While coefficient of correlation (r) indicates the strength of the relationship between two variables, coefficient of determination (r) actually measures that strength. Inflation Adjustment in a Time Series Analysis When working with a time series, especially one that covers a long period of time, the data can be distorted by the presence of inflation during that period. For instance, when working with sales growth over a ten-year period, the sales that occurred each year will contain an additional component representing the increase in the cost of that good simply as a result of inflation. We need a way to remove the portion of the sales each year that is attributable to price movements, leaving us with the change in the quantity of sales. This will provide us with inflation-adjusted sales figures and an inflation-adjusted sales forecast. 2, or .8984. Note that if the
Posted on: Fri, 01 Aug 2014 14:11:56 +0000
Trending Topics
Recently Viewed Topics
© 2015