Apparatus and method for identifying patterns in a multi-dimensional database

• US 7,676,468 B2
• Filed: 03/09/2007
• Issued: 03/09/2010
• Est. Priority Date: 03/28/2002
• Status: Active Grant

• Current Assignee: Business Objects Incorporated (SAP SE)
• Original Assignee: Business Objects Incorporated (SAP SE)
• Inventors: Macgregor, John Malcolm
• Primary Examiner(s): Le; Debbie M
• Application Number:
  US11/684,554
• Publication Number:
  US 20070150471A1
• Time in Patent Office: 1,096 Days
• Field of Search: 707 1- 7, 707100-1041
• US Class Current: 1/1
• CPC Class Codes: G06F 16/2462 Approximate or statistical queries
• CPC Class Codes: G06F 16/283 Multi-dimensional databases or data warehouses, e.g. MOLAP or ROLAP
• CPC Class Codes: G06F 2216/03 Data mining
• CPC Class Codes: Y10S 707/99936 Pattern matching access

Abstract

An apparatus to identify patterns in a multi-dimensional database is described. In particular, a computer-readable medium to direct a computer to function in a specified manner is provided. The computer-readable medium comprises instructions to automatically identify a plurality of patterns associated with data comprising the multi-dimensional database and instructions to indicate the plurality of patterns that are identified. Exemplary patterns that can be identified include an outlier pattern, a step pattern, a random pattern, a trend pattern, and a periodic pattern.

Claims

1. A computer-readable storage medium, comprising executable instructions to:

designate a plurality of patterns to apply to data within a multi-dimensional database, wherein the plurality of patterns include at least two different patterns selected from an outlier pattern, a step pattern, a random pattern, a trend pattern, and a periodic pattern; and

apply a plurality of statistical tests corresponding to the plurality of patterns to data vectors of the multi-dimensional database to identify at least one pattern, wherein the plurality of statistical tests are selected from Tukey'"'"'s test, a Standard Deviation test, a Runs test, an Autocorrelation test, a Mean Squared Successive Difference test, a Regression Analysis, and a Rank Correlation test.

2. The computer-readable storage medium of claim 1 further comprising executable instructions to display the at least one pattern.

3. A computer-implemented method, comprising:

defining a plurality of statistical tests selected from Tukey'"'"'s test, a Standard Deviation test, a Runs test, an Autocorrelation test, a Mean Squared Successive Difference test, a Regression Analysis, and a Rank Correlation test;

applying, with a processor, the plurality of statistical tests to a plurality of data vectors of a multi-dimensional database stored in memory to identify a plurality of patterns including at least two different patterns selected from an outlier pattern, a step pattern, a random pattern, a trend pattern, and a periodic pattern; and

displaying the plurality of patterns.

4. The computer-implemented method of claim 3, wherein the plurality of data vectors are oriented along a time dimension of the multi-dimensional database, and applying the plurality of statistical tests includes applying the plurality of statistical tests to the plurality of data vectors to identify the plurality of patterns with respect to the time dimension.