DATA DISAMBIGUATION SYSTEMS AND METHODS

• US 20100106675A1
• Filed: 07/20/2009
• Published: 04/29/2010
• Est. Priority Date: 05/04/2004
• Status: Active Grant

• Current Assignee: Verint Americas Incorporated (Verint Systems Incorporated)
• Original Assignee: Next IT Corporation (Verint Systems Incorporated)
• Inventors: Hust, Robert, Zartler, Mark
• Granted Patent: US 7,908,274 B2
• Time in Patent Office: Days
• US Class Current: 706/52
• CPC Class Codes: G06F 16/3329 Natural language query formulation or dialogue systems
• CPC Class Codes: G06F 16/3344 using natural language analysis
• CPC Class Codes: G06F 40/284 Lexical analysis, e.g. tokenisation or collocates
• CPC Class Codes: Y10S 707/99943 Generating database or data structure, e.g. via user interface
• CPC Class Codes: Y10S 707/99944 Object-oriented database structure

Abstract

Various embodiments provide a state-based, regular expression parser in which data, such as generally unstructured text, is received into the system and undergoes a tokenization process which permits structure to be imparted to the data. Tokenization of the data effectively enables various patterns in the data to be identified. In some embodiments, one or more components can utilize stimulus/response paradigms to recognize and react to patterns in the data.

Claims

1-103. -103. (canceled)

104. A computer-based system for determining a response to an input text string, the system comprising:

a server that receives the input text string via a computer data network, wherein the server executes software instructions stored on a computer readable medium, wherein the server is programmed to;

tokenize the input text string by parsing the input text string to define one or more recognizable patterns in the input text string;

comparing the one or more recognizable patterns to a plurality of cases of text to determine whether the one or more recognizable patterns match one or more of the plurality of cases, wherein each of the plurality of cases define a response to be taken in the event of a case match, wherein the cases of text are stored in a knowledge base and defined using a hierarchical tag-based markup language;

when a recognizable pattern matches only one case, perform the response for the case; and

when a recognizable pattern matches two or more cases;

score the two or more cases to determine the case with the highest probability match based on a scoring function; and

perform the response for the case with the highest probability match.

105. The system of claim 104, wherein the hierarchical tag-based markup language used to define the cases of text comprises:

an input tag that identifies a pattern of text to be matched; and

a response tag that identifies the response in the event of a case match.

106. The system of claim 104, wherein:

the plurality of cases of text comprises one or more input tags with exact text strings, one or more input tags with partial text string cases, and one or more input tags with variable text string cases; and

the scoring function scores exact text string case matches greater than variable text string case matches.

107. The system of claim 106, wherein:

the response tags for the cases define output text expressions to be output by the server; and

the hierarchical tag-based markup language used to define the cases of text further comprises;

a previous tag associated with an output expression, where the previous tag constrains a case from matching a recognizable pattern when the server did not previously output an output text expression that matches the output expression associated with the previous tag; and

a previous input tag associated with an input expression, where the previous input tag constrains a case from matching a recognizable pattern when the server did not previously receive from a user an input text expression that matches the input expression associated with the previous input tag.

108. The system of claim 104, further comprising a table stored in a computer readable medium, wherein the table is configured to contain a token assigned to the input text string.

109. The system of claim 104, further comprising:

a deployable agent, deployable by the server, comprising

a data source configured to provide a pipeline for data to travel; and

a runtime object configured to receive and process data that travels through the pipeline, wherein the data is text-based data.

110. The system of claim 109, wherein the deployable agent is an active agent configured to interact with a data origination entity.

111. The system of claim 109, wherein the deployable agent is a passive agent configured to receive information from a data origination entity.

112. A computer-based system for determining a response to an input text string, the system comprising:

a server that receives the input text string via a computer data network, wherein the server executes software instructions stored on a computer readable medium, wherein the server comprises;

one or more knowledge bases that store a plurality of cases of text, each case defining a response to be taken in the event of a case match, where the cases are defined using a hierarchical tag-based markup language; and

a functional presence engine that;

tokenizes the input text string by parsing the input text string to define one or more recognizable patterns in the input text string;

compares the one or more recognizable patterns to the plurality of cases of text in the one or more knowledge bases to determine whether the one or more recognizable patterns match one or more of the plurality of cases

when a recognizable pattern matches only one case, performs the response for the case; and

when a recognizable pattern matches two or more cases;

scores the two or more cases to determine the case with the highest probability match based on a scoring function; and

performs the response for the case with the highest probability match.

113. The system of claim 112, wherein the hierarchical tag-based markup language used to define the cases of text comprises:

an input tag that identifies a pattern of text to be matched; and

a response tag that identifies the response in the event of a case match.

114. The system of claim 112, wherein:

the plurality of cases of text comprises one or more input tags with exact text strings, one or more input tags with partial text string cases, and one or more input tags with variable text string cases; and

the scoring function scores exact text string case matches greater than variable text string case matches.

115. The system of claim 114, wherein:

the response tags for the cases define output text expressions to be output by the computer system; and

the hierarchical tag-based markup language used to define the cases of text further comprises;

a previous tag associated with an output expression, where the previous tag constrains a case from matching a recognizable pattern when the server did not previously output an output text expression that matches the output expression associated with the previous tag; and

a previous input tag associated with an input expression, where the previous input tag constrains a case from matching a recognizable pattern when the server did not previously receive from a user an input text expression that matches the input expression associated with the previous input tag.

116. The system of claim 112, further comprising a table stored in a computer readable medium, wherein the table is configured to contain a token assigned to the input text string by the functional presence engine.

117. The system of claim 112, further comprising:

a deployable agent, deployable by the server, comprising

a data source configured to provide a pipeline for data to travel; and

a runtime object configured to receive and process data that travels through the pipeline, wherein the data is text-based data.

118. The system of claim 117, wherein the deployable agent is an active agent configured to interact with a data origination entity.

119. The system of claim 117, wherein the deployable agent is a passive agent configured to receive information from a data origination entity.

120. A computer-implemented method comprising:

receiving a text string by a computer system;

tokenizing the text string by the computer system, wherein the tokenizing comprises parsing the input text string to define one or more recognizable patterns in the input text string;

comparing by the computer system the one or more recognizable patterns to a plurality of cases of text;

determining by the computer system whether the one or more recognizable patterns matches one or more of the plurality of cases, wherein each of the plurality of cases define a response to be taken in the event of a case match, wherein the cases of text are stored in a knowledge base and defined using a hierarchical tag-based markup language;

when a recognizable pattern matches only one case, performing the response for the case;

when a recognizable pattern matches two or more cases, scoring by the computer system the two or more cases to determine the case with the highest probability match based on a scoring function, and performing the response for the case with the highest probability match.

121. The computer-implemented method of claim 120, wherein the hierarchical tag-based markup language used to define the cases of text comprises:

an input tag that identifies a pattern of text to be matched; and

a response tag that identifies the response in the event of a case match.

122. The computer-implemented method of claim 120, wherein:

the plurality of cases of text comprises one or more input tags exact text strings, one or more input tags with partial text string cases, and one or more input tags with variable text string cases; and

for the scoring function, exact text string case matches are scored higher than variable text string case matches.

123. The computer-implemented method of claim 120, further comprising deploying by the computer system a deployable agent comprising a data source configured to provide a pipeline for data to travel and a runtime object configured to receive and process data that travels through the pipeline, wherein the data is text-based data.