Data mapping acceleration

• US 8,577,938 B2
• Filed: 08/23/2011
• Issued: 11/05/2013
• Est. Priority Date: 08/23/2011
• Status: Active Grant

• Current Assignee: Accenture Global Services Limited (Accenture PLC)
• Original Assignee: Accenture Global Services Limited (Accenture PLC)
• Inventors: Yeh, Peter Z., Verma, Kunal, Vasquez, Reymonrod, Srivatsa, Harsha Kumar, Kobe, Mark Gilbert
• Primary Examiner(s): AL HASHEMI, SANA A
• Application Number:
  US13/215,775
• Publication Number:
  US 20130054605A1
• Time in Patent Office: 805 Days
• Field of Search: 707/687, 707/705, 707/709, 707/813, 707/821, 707/964
• US Class Current: 707/821
• CPC Class Codes: G06F 16/258 Data format conversion from or to a database

Abstract

A data mapping acceleration system may include a syntactic profile module to generate a syntactic profile of a data source to be mapped. A semantic classification module may generate a semantic classification of the data source. A similarity module may evaluate the syntactic profile and semantic classification to determine an overall similarity between attributes of data from the data source for mapping the data source.

Claims

1. A data mapping acceleration system comprising:

a syntactic profile module, executed by a processor, to generate a syntactic profile of a data source to be mapped;

a semantic classification module to generate a semantic classification of the data source by classifying values for a column of data from the data source by using an ontology that includes hierarchal classes, wherein a class of the hierarchal classes includes an attribute of including a lexical realization, and using classes from the ontology for which the lexical realization of the classes matches a column name of the column of the data; and

a similarity module to evaluate the syntactic profile and semantic classification to determine an overall similarity between attributes of the data from the data source for mapping the data source.

4. The system of claim 1, wherein the syntactic profile includes a token vector for a column of the data from the data source, and wherein the token vector includes the data separated from an original format into a tokenized format including at least one of words, phrases, and symbols.

5. The system of claim 1, wherein the overall similarity includes a component of a syntactic similarity and a semantic similarity.

6. The system of claim 1, wherein the class of the hierarchal classes further includes the attribute of including a classification type that is selected from pattern-based, keyword-based, rule-based, and classifier-based classification types.

7. The system of claim 1, wherein the similarity module is to determine a value representing the overall similarity by:

determining a syntactic similarity by evaluating;

tokens from a token vector for an attribute of the attributes;

unique non-null tokens from the token vector for the attribute; and

a number of times a token occurs in the attribute according to the token vector for the attribute.

8. The system of claim 1, wherein the similarity module is to determine a value representing the overall similarity by:

determining a semantic similarity by evaluating;

unique classes from a class vector for an attribute of the attributes; and

a number of times a class occurs in the attribute according to the class vector for the attribute.

9. The system of claim 1, wherein the similarity module is to determine a value representing the overall similarity by:

applying a discount factor to a syntactic similarity based on a number of null values per column of the data; and

applying another discount factor to a semantic similarity based a number of classes for a particular column of the data.

15. The system of claim 6, wherein the semantic classification module is to:

use the pattern-based classification to generate the semantic classification based on a pattern of values;

use the keyword-based classification to generate the semantic classification based on a comparison of keywords to values of the column of the data to determine whether a value matches a keyword;

use the rule-based classification to generate the semantic classification based on a classification of values of the column of the data based on pre-defined rules; and

use the classifier-based classification to generate the semantic classification based on a classification of values of the column of the data based on previously determined classifications to determine a match.

2. A method for data mapping acceleration, the method comprising:

generating, by a processor, a syntactic profile of a data source to be mapped;

generating a semantic classification of the data source by classifying values for a column of data from the data source by using an ontology that includes hierarchal classes, wherein a class of the hierarchal classes includes an attribute of including a lexical realization, and using classes from the ontology for which the lexical realization of the classes matches a column name of the column of the data; and

evaluating the syntactic profile and semantic classification to determine an overall similarity between attributes of the data from the data source for mapping the data source.

10. The method of claim 2, further comprising tokenizing values for a column of the data from the data source to determine the syntactic profile.

11. The method of claim 2, further comprising generating a token vector for a column of the data from the data source to determine the syntactic profile.

12. The method of claim 2, further comprising generating a class vector for a column of the data from the data source to determine the semantic classification.

13. The method of claim 2, wherein if a value of a column of the data from the data source includes multiple classifications that include a sub-class and super class relationship, the method further comprising identifying subsumption relationships and retaining a specific subsumee for determining the semantic classification.

14. The method of claim 2, wherein if a value of the column of the data from the data source cannot be classified, the method further comprising:

identifying a class of the ontology that constitutes a majority; and

using the identified class to classify the value.

16. The method of claim 10, wherein the tokenized values are designated tokens, the method further comprising using a glossary for normalizing the tokens.

3. A non-transitory computer readable medium having stored thereon a computer executable program to provide data mapping acceleration, the computer executable program when executed causes a computer system to:

generate, by a processor, a syntactic profile of a data source to be mapped;

generate a semantic classification of the data source by classifying values for a column of data from the data source by using an ontology that includes hierarchal classes, wherein a class of the hierarchal classes includes an attribute of including a lexical realization, and using classes from the ontology for which the lexical realization of the classes matches a column name of the column of the data; and

evaluate the syntactic profile and semantic classification to determine an overall similarity between attributes of the data from the data source for mapping the data source.