Judgmental Models Essay Example
Judgmental Models Free Essay Example
Judgmental Models – These are set of techniques that include Delphi Model, that rely on expertise, experience and intuition. These are some of the historical data that were not available or when the decision makers needed to foresee further into the future.
Statistical Time series models are mostly applicable for short range forecasting problems. The Statistical time series model assumes that whatever forces have influenced sales in the recent past will continue into the future. Time series forecast are basically developed when trying extrapolating the data for the future.
Explanatory / casual model is a model that seeks at identifying the factors that explain statistically the given patterns observed in the variable being forecast, it usually integrated with the regression analysis. The main difference includes long term nature of the judgmental model vs. short term model; as well as implementing exploratory/causal models that includes factors other than time when trying to explain the changes in the dependent variable.
There are two judgmental forecasting approaches discussed in the chapter, The Historical analogy and the Delphi Method. Historical analogy- is a forecast that is obtained through a comparative analysis with a previous situation, with careful attention being paid in observing the different trends and ways that the business environment is changing; thus providing the management team with an opportunity of checking and predicting the future of the business.
Delphi method–uses a panel of experts that are identified by the organization and their identity are kept as a typical secret from each other this is basically in the process of filling the questioners. Thus after each round of responses the results that are delivered by the experts are edited to ensure that there ideas are shared thus it will provide them the opportunity of knowing what each expert is thinking on what the others and this process is repeated to see if the experts are thinking about a given business issues, Delphi method promotes unbiased exchange if ideas and discussion and this results in some convergence of opinions.
Mean Absolute Deviance (MAD)- this is the absolute difference between the actual value and the forecasts, average over a range of forecasted values. Accurate and Robust measure of forcast accuracy not sensitive to outliers.
Penalizes all the deviation consistently (Linearly). Mean Square Error (MSE) – this is the squared difference between the actual value and the forecast, averaged over forecasted values. Penalizes outliers diasporiatinally more (quadratically), thus more sensitive to extreme outliers. Mean Absolute Percentage Error (MAPE) this is known as the average sum of absolute errors divided by actual observation values. Since the deviation are divided by actual observations. The measure is relative and unit less where as the previous two mainly depend on the units of observations and are seen as absolute in that sense.
Exponential smoothing – this is the next period forecast equals the weighted combination of the actual observations and the previous period forecast. Alternatively Exponential Smoothing is expressed as the previous period forecasting in addition with the fraction that is found in the forecasted error. In either way, by continually substituting the previous forecasts, all these result to the beginning of time horizon, we get that the exponential smoothing forecast is just weighted moving average, where weights exponentially diminish in the further back into the past.
These are variables measured based on a continuous measurement scale in the form of (Length, Time, and weights amongst others) while the attributes are resulted from counting (visual inspection, good-no good, etc)
In control refers to having the only common causes of variation present in a given business process.
Selecting a sample of observations from a production or service process, this is measuring one or more quality characteristics, making few calculations, recording data, plotting key statistics on a control chart, examining the chart to determine if any unusual patterns are reached. This is called out of control condition identified this will determine the caused out of control condition and taking corrective action.