Technique for determination of an exception in multi-dimensional data

1. A method for determination of an exception in multi-dimensional data, the method comprising the steps of:

a) determining an expected value for a set of cells of the multi-dimensional data by means of an ANOVA based method;

b) determining a residual for each cell of the set of cells;

c) scaling the residuals; and

d) comparing the scaled residuals with a threshold value for determination of the exception.

2. The method of claim 1 , wherein the multi-dimensional data is stored in an OLAP data base.

3. The method of claim 1 , further comprising the step of performing a variance stabilization transformation prior to said comparing step.

4. The method of claim 3 , wherein the variance stabilization transformation comprises a non-linear transformation function.

5. The method of claim 3 , wherein the variance stabilization transformation comprises reducing weight of large absolute data values with respect to low absolute data values of the multi-dimensional data.

6. The method of claim 3 , wherein the variance stabilization transformation comprises a transformation function having a slope greater than a constant for a data interval comprising 0 and having a slope lesser than the constant outside the data interval.

7. The method of claim 3 , wherein the variance stabilization transformation is performed with respect to all data values or a selected sub-set of data values.

8. The method of claim 1 , wherein the ANOVA based method is a linear ANOVA based method.

9. The method of claim 1 , wherein the residual for each cell is calculated by subtracting an expected value from an observed value of each cell.

10. The method of claims 1 , wherein said scaling step further comprises calculating the standard deviation of the residuals from all cells of a same aggregation level and standardizing the residual of the same aggregation level by dividing the residual by the associated standard deviation.

11. The method of claim 10 , wherein the exception is determined within the same aggregation level.

12. The method of claim 11 , wherein the aggregation level is user specified.

13. A computer program product stored on a computer usable medium comprising computer readable program means for determination of an exception in multi-dimensional data, said computer program product comprising:

first subprocesses for determining an expected value for a set of cells of the multi-dimensional data by means of an ANOVA based method,

second subprocesses for determining a residual for each cell of the set of cells;

third subprocesses for scaling the residuals; and

fouth subprocessess for sucomparing the scaled residuals with a threshold value for determination of the exception.

14. A computer system for determination of an exception in multi-dimensional data, said system comprising:

means for determining an expected value for a set of cells of the multi-dimensional data by means of an ANOVA based method;

means for determining a residual for each cell of the set of cells;

means for scaling the residuals; and

means for comparing the scaled residuals with a threshold value for determination of the exception.