Hr Inc Essay

293 WordsNov 4, 20122 Pages
Human Resources, Inc. We analyze this case using two different approaches: (1) multiplicative decomposition model on the time-series data, and (2) regression model. Multiplicative decomposition model. The results of this analysis are shown in sheet named Decomp in file P11- Human.XLS. The error graph is also shown. We have time-series data for 36 quarters and there are 4 seasons each year. The resulting MAPE is 8.58%. The error graph indicates that while the forecasting model does a good job overall of replicating the pattern of participant values, there are sizable differences in some quarters. The relatively large errors in the last two quarters are particularly disturbing and could indicate, for example, some recent occurrence (such as a new competitor) that is obviously not being detected by the model. Regression model. A simple regression model would be to use only the quarter number as the independent variable. However, since we expect the seasons (and locations) to have an impact on participation, we need to expand the independent variable set to include the seasons. To do so, we include a variable for three of the four seasons. We do not include a variable for the fourth season since this would make the columns linearly dependent. The resulting equation (see sheet named Regression in file P11-Human.XLS) is: Participants  54.028  1.238 Quarter number  0.824 Winter  9.081 Spring  1.015 Fall The coefficient of determination is 78.8% and the standard error is 7.611. Students should be asked to interpret these values. The MAPE is 8.79%, which is slightly more than the MAPE obtained using the decomposition model. The error graph reveals a similar pattern to the earlier model, with sizable differences in some