Question

What are some of the criteria for forecast evaluation? What is Theil’s statistic?

Answer 1

Some of the criteria are as follows:

1. Mean absolute deviation (MAD). It is used to identify the average of the absolute deviation between the actual and forecasted results. A smaller value of MAD, means higher accuracy in the forecast and vice versa.

2. Mean Square error (MSE). It is identified by calculating the mean of the square of the deviations between the actual and forecasted data. Here, the square is done to maximize the impact of deviation and make evaluation to be more accurate. A lower value of MSE is interpreted as relatively accurate forecasting and vice versa.

3. Mean absolute percentage error (MAPE). It is the method where, mean of the absolute deviation as a percentage of actual data is identified and interpreted. A higher MAPE value, means higher deviation and inaccuracy in the results and vice versa.

Theil statistics is a technique that is used to measure the relative accuracy of the forecsted data. Here, one forecasted result is identified using the traditional method and another forecasted result is identified using minimum or no any historical data. Now, these two forecasted data are compared using the formula suggested by Theil statistics and relative accuracy is identified.