System and method for computerized batching of huge populations of electronic documents

• US 9,760,622 B2
• Filed: 02/27/2015
• Issued: 09/12/2017
• Est. Priority Date: 08/08/2012
• Status: Active Grant

• Current Assignee: Microsoft Israel Research And Development (2002) Ltd (Microsoft Corporation)
• Original Assignee: Microsoft Israel Research And Development (2002) Ltd (Microsoft Corporation)
• Inventors: Ravid, Yiftach
• Primary Examiner(s): Arjomandi, Noosha
• Application Number:
  US14/633,906
• Publication Number:
  US 20160034556A1
• Time in Patent Office: 928 Days
• Field of Search: 707737
• CPC Class Codes: G06F 16/285 Clustering or classification
• CPC Class Codes: G06F 16/355 Class or cluster creation or modification
• CPC Class Codes: G06F 16/93 Document management systems

Abstract

A method for computerized batching of huge populations of electronic documents, including computerized assignment of electronic documents into at least one sequence of electronic document batches such that each document is assigned to a batch in the sequence of batches and such that there is no conflict between batching requirements, the following batching requirements being maintained by a suitably programmed processor: a. pre-defined subsets of documents are always kept together in the same batch, b. batches are equal in size, c. the population is partitioned into clusters, and all documents in any given batch belong to a single cluster rather than to two or more clusters.

Claims

1. A method for computerized batching of a population of

electronic documents, comprising assigning the electronic documents into at least one sequence of batches such that each electronic document is assigned to a batch in the sequence of batches and such that the assigning attempts to avoid conflict between batching requirements, the batching requirements comprising:

a. pre-defined subsets of the electronic documents are always kept together in a same batch, the electronic documents in a pre-defined subset are assigned a same keep-together value,

b. the batches in the sequence of batches are equal in size, the size being a measurable batch metric, wherein batching requirement that pre-defined subsets of the electronic documents are always kept together in the same batch takes precedence over the batching requirement that the batched are equal in size such that documents with the same keep-together value are assigned to a single batch regardless of a resulting size of the batch, and

c. the population of electronic documents is partitioned into clusters and all electronic documents in any given batch belong to a single cluster rather than to two or more clusters.

2. The method according to claim 1, wherein the batching requirements further comprise:

positions of individual electronic documents within the sequence of batches are determined by a pre-known urgency for review such that documents of higher relevance appear in early batches and documents of lower relevance appear in later batches.

3. The method according to claim 1, further comprising computing an urgency score for at least one keep-together subset which represents an attribute associated with a central tendency of urgency scores of all the electronic documents in the keep-together subset.

4. The method according to claim 3, wherein the urgency score comprises a most frequently occurring urgency score of all the electronic documents in the keep-together subset.

5. The method according to claim 1, further comprising:

filling batches, separately for each of the clusters, including using first keep-together subsets as batches and then combining second keep-together subsets other than the first keep-together subsets into batches; and

ordering the batches into the sequence according to urgency.

6. The method according to claim 1, further comprising:

filling batches, separately for each of the clusters, including combining first keep-together subsets other than small keep-together subsets into batches; and

subsequently passing over the batches in an order determined by known urgency and enlarging at least some of the batches by adding the small keep-together subsets thereto, in the order determined by the known urgency.

7. The method according to claim 6, wherein a small keep-together subset comprises only a single document.

8. The method according to claim 1, wherein the batching requirement that the population of electronic documents is partitioned into clusters and all electronic documents in any given batch belong to a single cluster rather than to two or more clusters is based on artificial intelligence functionality that uniformizes cluster identifications of all electronic documents having a single keep-together identification.

9. The method according to claim 1 wherein due to conflict between requirements (b), (c), there is for at least one duster, a batch of lesser size than other batches equal in size.

10. The method according to claim 1 wherein, due to conflict between batching requirements (a) and (b), there is at least one batch of greater size than other batches equal size.

11. A non-transitory computer readable medium having computer readable program code embodied therein, said computer readable program code executable to implement the method of claim 1.

12. A method for computerized batching of a population of

electronic documents, comprising assigning the electronic documents into at least one sequence of batches such that each electronic document is assigned to a batch in the sequence of batches and such that the assigning attempts to avoid conflict between batching requirements, the batching requirements comprising:

a. pre-defined subsets of the electronic documents are always kept together in a same batch, the electronic documents in a pre-defined subset are assigned a same keep-together value,

b. the batches in the sequence of batches are equal in size, the size being a measurable batch metric, wherein batching requirement that pre-defined subsets of the electronic documents are always kept together in the same batch takes precedence over the batching require lent that the batched are equal in size such that documents with the same keep-together value are assigned to a single batch regardless of a resulting size of the batch, and

c. positions of individual electronic documents within the sequence of batches are determined by a pre-known urgency for review such that documents of higher relevance appear in early batches and documents of lower relevance appear in later batches.

13. The method according to claim 12 wherein, due to conflict between batching requirements (a) and (b), there is at least one batch of greater size than other batches equal size.

14. The method according to claim 12, further comprising computing an urgency score for at least one keep-together subset which represents an attribute associated with a central tendency of urgency scores of all the electronic documents in the keep-together subset.

15. The method according to claim 14, wherein the urgency score comprises a most frequently occurring urgency score of all the electronic documents in the keep-together subset.

16. The method according to claim 12, further comprising:

filling batches, separately for each clusters, including using first keep-together subsets as batches and then combining second keep-together subsets other than the first keep-together subsets into batches; and

ordering the batches into the sequence according to urgency.

17. The method according to claim 12, further comprising:

filling batches, separately for each clusters, including combining first keep-together sets other than small keep-together subsets into batches; and

subsequently passing over the batches in an order determined by known urgency and enlarging at least some of the batches by adding the small keep-together subsets thereto, in the order determined by the known urgency.

18. The method according to claim 17, wherein a small keep-together subset comprises only a single document.

19. The method according to claim 12 wherein the batching requirement that the population of electronic documents is partitioned into clusters and all electronic documents in any given batch belong to a single duster rather than to two or more clusters is based on artificial intelligence functionality that uniformizes cluster identifications of all electronic documents having a single keep-together identification.

20. A non-transitory computer readable medium having computer readable program code embodied therein, said computer readable program code executable to implement the method of claim 12.

21. A system for computerized batching of a population of electronic documents, the system including:

a processor for controlling computerized assignment of the electronic documents into at least one sequence of batches such that each electronic document is assigned to a batch in the sequence of batches and such that the computerized assignment attempts to avoid conflict between batching requirements, the batching requirements comprising;

a. pre-defined subsets of the electronic documents are always kept together in a same hatch, the electronic documents in a pre-defined subset are assigned a same keep-together value,

b. the batches in the sequence of batches are equal in size, the size being a measurable batch metric, and

c. the population of electronic documents is partitioned into clusters and all electronic documents in any given batch belong to a single cluster rather than to two or more clusters.

22. A system for computerized batching of a population of electronic documents, the system including:

a processor for controlling computerized assignment of the electronic documents into at least one sequence of batches such that each electronic document is assigned to a batch in the sequence of batches and such that the computerized assignment attempts to avoid conflict between batching requirements, the batching requirements comprising;

a. pre-defined subsets of the electronic documents are always kept together in a same batch, the electronic documents in a pre-defined subset are assigned a same keep-together value,

b. the batches in the sequence of batches are equal in size, the size being a measurable batch metric, and

c. positions of individual electronic documents within the sequence of batches are determined by a pre-known urgency for review such that documents of higher relevance appear in early batches and documents of lower relevance appear in later batches.