INFERRING RULES TO CLASSIFY OBJECTS IN A FILE MANAGEMENT SYSTEM

• US 20100161621A1
• Filed: 12/19/2008
• Published: 06/24/2010
• Est. Priority Date: 12/19/2008
• Status: Active Grant

• Current Assignee: SAP SE
• Original Assignee: SAP AG (SAP SE)
• Inventors: PETERS, JOHAN CHRISTIAAN
• Granted Patent: US 8,099,419 B2
• Time in Patent Office: Days
• US Class Current: 707/749
• CPC Class Codes: G06F 16/16 File or folder operations, e.g. details of user interfaces specifically adapted to file systems

Abstract

Described are methods and systems related to inferring rules to classify an object to one of one or more target folders. One or more properties of the object to be classified are determined. A first degree of affinity between the object to be classified and the objects of the target folders, having a property identical to the object to be classified, is computed. A second degree of affinity between the object to be classified and the objects within each target folder, having a property identical to the object to be classified, is computed. A total degree of affinity between the object to be classified and each target folder is calculated. A normalized total degree of affinity is calculated by averaging the total degree of affinity across all target folders. The object is moved to a target folder having a highest value of the normalized total degree of affinity.

Claims

1. An article of manufacture, comprising a machine readable medium having instructions which when executed by a machine cause the machine to perform operations comprising:

determining one or more properties of the object to be classified;

computing a first degree of affinity for each property common between the object to be classified and objects of the target folder, the first degree of affinity indicative of an affinity of the object to each target folder and a uniqueness of the affinity across all target folders;

computing a second degree of affinity for each property common between the object to be classified and the objects of the target folder, the second degree of affinity indicative of an affinity of the object within each target folder and a statistical significance of the affinity within the target folder;

calculating a total degree of affinity between the object to be classified and each target folder, the total degree of affinity being a sum of a product of the first degree of affinity and the second degree of affinity for all properties of the object;

calculating a normalized total degree of affinity, the normalized total degree of affinity calculated by averaging the total degree of affinity across all target folders;

identifying at least one of the target folders as having a highest value of the normalized total degree of affinity; and

moving the object to be classified to the identified folder.

2. The article of manufacture of claim 1, wherein computing the first degree of affinity comprises:

for each target folder, determining properties common between the object to be classified and objects of the target folder;

for each common property, determining a per folder number of objects of the target folder having property identical to the object to be classified;

determining a total number of objects across all target folders having a property identical to the object to be classified; and

calculating a ratio between the per folder number of objects and the total number of objects.

3. The article of manufacture of claim 1, wherein computing a second degree of affinity comprises:

selecting a property of the object to be classified;

for each target folder, determining a number of objects having a property identical the object to be classified;

determining a total number of objects in corresponding target folders; and

calculating a ratio between the number of objects with identical properties and the total number of objects.

4. A computer implemented method for inferring rules to classify an object to one of one or more target folders, the method comprising:

determining one or more properties of the object to be classified to one of the target folders, each target folder having one or more objects and corresponding one or more properties of the object;

computing a between-folder affinity for each property common between the object to be classified and the objects of the target folder, by

determining a number of objects having a property identical to the object to be classified, within each target folder,

determining a total number of objects having a property identical to the object to be classified, across all target folders, and

calculating a ratio between the number of objects with identical properties and the total number of objects;

computing an in-folder affinity for each property common between the object to be classified and the objects of the target folder, by

determining a number of objects having a property identical to the object to be classified, within each target folder,

determining a total number of objects in corresponding target folder, and

calculating a ratio between the number of objects with identical properties and the total number of objects;

calculating a total degree of affinity between the object to be classified and each target folder, by

determining a product of the between-folder affinity and the in-folder affinity for each property of the objects in each target folder, and

determining a sum of the product of between-folder affinity and the in-folder affinity for each target folder;

calculating a normalized total degree of affinity for each target folder, by

determining a sum of the total degree of affinity of all the target folders, and

determining a ratio of the total degree of affinity for each target folder and the sum of the total degree of affinity of all target folders; and

moving the object to a target folder having a highest value of the normalized total degree of affinity.

5. The computer implemented method of claim 4, wherein the object to be classified, may be residing in a source folder.

6. The computer implemented method of claim 4, wherein the object is classified as belonging to one of the target folders based on a function of the first degree of affinity and the second degree of affinity.

7. The computer implemented method of claim 4, wherein, if there are more than one target folders having the highest value of the normalized total degree of affinity, providing an option to choose the target folder to move the object.

8. The computer implemented method of claim 4, wherein based on the inferred rules one or more objects may be categorized to a respective target folder.

9. The computer implemented method of claim 4, wherein the inferred rules may be stored as a template.

10. The computer implemented method of claim 4, wherein a view of one of the one or more target folders may be generated based upon the inferred rules.

11. The computer implemented method of claim 4, wherein a user action performed on one of the one or more target folders may generate a tooltip to view information about the one or more objects already present in corresponding target folder.

12. The computer implemented method of claim 4, wherein the target folders are indexed based upon a precedence of objects stored therein.

13. The computer implemented method of claim 4, further comprising moving the object to a target folder based upon inferred rules stored in an index.

14. The computer implemented method of claim 13, wherein the index is updated based upon one or more changes that occur when an object is moved to a target folder.

15. The computer implemented method of claim 14, wherein, based upon the changes that occur when an object is moved to a target folder, the inferred rules stored in the index are updated.

17. The system of claim 14, further comprising an index in communication with the target folder repository, the metadata repository, the metadata reader, the affinity calculator and the object classifying engine, for storing the inferred rules along with parameters of the rules.

18. The computer implemented method of claim 15, wherein the index further comprises:

one or more properties of the object to be classified to one of the target folders;

one or more objects present in each target folder and corresponding one or more properties describing the object;

a between-folder affinity of each property of the objects present in each target folder,

an in-folder affinity of each property of the objects present in each target folder,

a total degree of affinity for each property of the objects present in each target folder,

a normalized total affinity for each target folder, and

a highest value of normalized affinity.

20. The system of claim 14, further comprises a memory for storing the inferred rules.

16. A system for inferring rules to classify an object to one of one or more target folders, the system comprising:

a target folder repository for storing the target folders, the target folders having one or more objects and corresponding one or more properties describing the object;

a metadata repository in communication with the target folder repository, for storing information about the target folders and contents of the target folders;

a metadata reader in communication with the metadata repository, for determining one or more properties of the object to be classified, determining a total degree of affinity for each property of the objects present in each target folder, determining a normalized total affinity for each target folder and determining a target folder having a highest value of the normalized total affinity;

an affinity calculator in communication with the metadata repository and the metadata reader, for calculating a between-folder affinity of each property of the objects present in each target folder, and calculating an in-folder affinity of each property of the objects present in each target folder; and

an object classifying engine in communication with the affinity calculator, the metadata reader and the target folder repository, for moving the object to a target folder having a highest value of the normalized affinity.

19. The method of claim 16, wherein the index is updated based upon a modification in contents of the index.