Error checking database records

• US 10,204,120 B2
• Filed: 09/19/2014
• Issued: 02/12/2019
• Est. Priority Date: 09/19/2014
• Status: Active Grant

• Current Assignee: Salesforce.com, Inc.
• Original Assignee: Salesforce.com, Inc.
• Inventors: Fuchs, Matthew
• Primary Examiner(s): Phillips, III, Albert M
• Application Number:
  US14/490,841
• Publication Number:
  US 20160085789A1
• Time in Patent Office: 1,607 Days
• CPC Class Codes: G06F 16/215 Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
• CPC Class Codes: G06F 16/2365 Ensuring data consistency and integrity

Abstract

An error checking technique for database records. A record is selected and its entities are compared with the entities of other records stored in the database to determine a likelihood that the labels associated with the entities of the selected record are correct. The likelihood for each entity of the selected record being correctly labeled can be determined by comparing the number of times that the entity appears in the database records with that label to the number of times that the entity appears in the database records with any other label. If the likelihood does not exceed a threshold, then an error is likely, and action can be taken to correct the record.

Claims

1. An apparatus for performing a quality check on database records, the apparatus comprising:

a processor; and

a non-transitory computer readable medium storing a plurality of instructions, which when executed, cause the one or more processors to;

select, by a database system, a first record comprising a first entity being assigned a first predefined label, and a second entity being assigned a second predefined label, the second predefined label being different than the first predefined label, the first entity including multiple terms;

rearrange, by the database system, the multiple terms of the first entity into a plurality of permutations;

evaluate, by the database system, for each permutation, a likelihood that the permutation corresponds to the first entity of the first record;

determine, by the database system, a first number of times that the permutation of the first entity with a highest likelihood is assigned to the first predefined label in stored database records;

determine, by the database system, a second number of times that the permutation of the first entity with the highest likelihood is assigned to any label in the stored database records;

determine, by the database system, a likelihood that the first predefined label that is assigned to permutation of the first entity with the highest likelihood is correct, wherein the assigned first predefined label likelihood is determined based the first number and the second number; and

initiate, by the database system, action to correct the first record when the assigned first predefined label likelihood is less than a first threshold.

2. The apparatus of claim 1, wherein the instructions further cause the processor to:

compare a first entity of the first record with the corresponding first entity of the stored database records to determine a first frequency of times that the first predefined label of the first entity of the first record is assigned to the corresponding first entity of the stored database records;

compare the first entity of the first record with the other entities of the stored database records to determine a second frequency of times that the first predefined label of the first entity of the first record is assigned to any entity of the stored database records; and

initiate action to correct the first predefined label of the first entity of the first record when the ratio of the first frequency to the second frequency is less than a threshold.

3. The apparatus of claim 2, wherein the instructions further cause the processor to:

perform the steps to compare the first entity for each of the multiple terms when the first entity includes multiple terms.

4. The apparatus of claim 1, wherein the instructions further cause the processor to:

compare a plurality of the entities of the first record with the corresponding entities of the stored database records to determine a first frequency of times that the labels assigned to the entities of the first record match the corresponding entities of the stored database records;

compare the plurality of the entities of the first record with all other entities of the stored database records to determine a second frequency of times that the labels assigned to the entities of the first record match the entities of any of the stored database records;

initiate action to correct at least one label of at least one entity of the first record when the ratio of the first frequency to the second frequency is less than a threshold.

5. The apparatus of claim 4, wherein the instructions further, cause the processor to:

determine the likelihood that all of the labels of the first record are correct by calculating for each entity of the first record the probability that the entity is correct, and multiplying together the probabilities for all entities of the first record.

6. A computer program product comprising computer-readable program code to be executed by one or more processors when retrieved from a non-transitory computer-readable medium, the program code including instructions to:

select, by a database system, a first record comprising a first entity being assigned a first predefined label and a second entity being assigned a second predefined label, the second predefined label being different than the first predefined label, the first entity including multiple terms;

rearrange, by the database system, the multiple terms of the first entity into a plurality of permutations;

evaluate, by the database system, for each permutation, a likelihood that the permutation corresponds to the first entity of the first record;

determine, by the database system, a first number of times that the permutation of the first entity with a highest likelihood is assigned to the first predefined label in stored database records;

determine, by the database system, a second number of times that the permutation of the first entity with the highest likelihood is assigned to any label in the stored database records;

determine, by the database system, a likelihood that the first predefined label that is assigned to permutation of the first entity with the highest likelihood is correct, wherein the assigned first predefined label likelihood is determined based the first number and the second number; and

initiate, by the database system, action to correct the first record when the assigned first predefined label likelihood is less than a first threshold.

7. The computer program product of claim 6, wherein the program code includes further instructions to:

compare a first entity of the first record with the corresponding first entity of the stored database records to determine a first frequency of times that the first predefined label of the first entity of the first record is assigned to the corresponding first entity of the stored database records;

compare the first entity of the first record with the other entities of the stored database records to determine a second frequency of times that the first predefined label of the first entity of the first record is assigned to any entity of the stored database records; and

initiate action to correct the first predefined label of the first entity of the first record when the ratio of the first frequency to the second frequency is less than a threshold.

8. The computer program product of claim 7, wherein the program code includes further instructions to:

perform the steps to compare the first entity for each of the multiple terms when the first entity includes multiple terms.

9. The computer program product of claim 6, wherein the program code includes further instructions to:

compare a plurality of the entities of the first record with the corresponding entities of the stored database records to determine a first frequency of times that the labels assigned to the entities of the first record match the corresponding entities of the stored database records;

compare the plurality of the entities of the first record with all other entities of the stored database records to determine a second frequency of times that the labels assigned to the entities of the first record match the entities of any of the stored database records;

initiate action to correct at least one label of at least one entity of the first record when the ratio of the first frequency to the second frequency is less than a threshold.

10. A method for performing a quality check on database records, the method comprising:

selecting, by a database system, a first record comprising a first entity being assigned a first predefined label and a second entity being assigned a second predefined label, the second predefined label being different than the first predefined label, the first entity including multiple terms;

rearranging, by the database system, the multiple terms of the first entity into a plurality of permutations;

evaluating, by the database system, for each permutation, a likelihood that the permutation corresponds to the first entity of the first record,

determining, by the database system, a first number of times that the permutation of the first entity with a highest likelihood is assigned to the first predefined label in stored database records;

determining, by the database system, a second number of times that the permutation of the first entity with the highest likelihood is assigned to any label in the stored database records;

determining, by the database system, a likelihood that the first predefined label that is assigned to permutation of the first entity with the highest likelihood is correct, wherein the assigned first predefined label likelihood is determined based the first number and the second number; and

initiating, by the database system, action to correct the first record when the assigned first predefined label likelihood is less than a first threshold.

11. The method of claim 10, further comprising:

comparing a first entity of the first record with the corresponding first entity of the stored database records to determine a first frequency of times that the first predefined label of, the first entity of the first record is assigned to the corresponding first entity of the stored database records;

comparing the first entity of the first record with the other entities of the stored database records to determine a second frequency of times that the first predefined label of the first entity of the first record is assigned to any entity of the stored database records; and

when the ratio of the first frequency to the second frequency is less than a threshold, initiate action to correct the first predefined label of the first entity of the first record.

12. The method of claim 11, further comprising:

when the first entity includes multiple terms, performing the steps to compare the first entity for each of the multiple terms.

13. The method of claim 10, further comprising:

comparing a plurality of the entities of the first record with the corresponding entities of the stored database records to determine a first frequency of times that the labels assigned to the entities of the first record match the corresponding entities of the stored database records;

comparing the plurality of the entities of the first record with all other entities of the stored database records to determine a second frequency of times that the labels assigned to the entities of the first record match the entities of any of the stored database records;

when the ratio of the first frequency to the second frequency is less than a threshold, initiate action to correct at least one label of at least one entity of the first record.

14. The method of claim 13, further comprising:

determine the likelihood that all of the labels of the first record are correct by calculating for each entity of the first record the probability that the entity is correct, and multiplying together the probabilities for all entities of the first record.