Methods and systems for template driven data mining task editing

• US 8,370,386 B1
• Filed: 11/03/2009
• Issued: 02/05/2013
• Est. Priority Date: 11/03/2009
• Status: Active Grant

• Current Assignee: The Boeing Co.
• Original Assignee: The Boeing Co.
• Inventors: Wang, Changzhou, Choi, Jai Joon
• Primary Examiner(s): Lee, Wilson
• Assistant Examiner(s): HERSHLEY, MARK E
• Application Number:
  US12/611,624
• Time in Patent Office: 1,190 Days
• Field of Search: 707/776, 707/778, 707/779
• US Class Current: 707/778
• CPC Class Codes: G06F 16/9038 Presentation of query results
• CPC Class Codes: G06F 2216/03 Data mining

Abstract

A computer-based method for creating data mining task templates for utilization in data mining activities is described. The method includes defining, via a template editor and an associated user interface, a task template for discovery of common patterns occurring within data mining events, generating a task editor/wizard from the task template, creating example tasks via the task template and task editor/wizard, running at least one of the example tasks, and refining the task template using results returned from execution of the at least one of the example tasks.

Claims

1. A computer-based method for creating data mining task templates for utilization in data mining activities, said method comprising:

receiving a task template for discovery of common patterns occurring within data mining events;

creating example tasks from the task template by interpreting the task template with specific values of variables, wherein interpreting the task template comprises;

creating a root element of a task from a root element associated with the task template when the variables are a collection data type; and

generating attributes and child nodes of the task using the task root element and the task template root element, wherein generating attributes and child nodes of the task using the task root element and the task root template element comprises;

looping through all the attributes of the task template root element;

obtaining a list of text strings from a template string specified as a value of an attribute for each selected attribute in the task template root element;

creating attributes using the task root element, an attribute name specified by an attribute in the task template root element, and the obtained list of text strings from the same attribute in the task template root element for each selected attribute in the task template root element; and

creating child elements and text nodes using the task root element and the task template root element;

running at least one of the example tasks; and

refining the task template using results returned from execution of the at least one of the example tasks.

2. The method according to claim 1 wherein interpreting the task template with variable values comprises:

replacing the variables with string representations with attribute values and element text contents;

removing attributes and elements if a corresponding select element has a value of ‘

false’

; and

replacing each element with a corresponding child when the corresponding select element has a value of ‘

children’

.

3. The method according to claim 1 wherein creating child elements and text nodes using the task root element and the task template root element comprises:

looping through all the child nodes of the task template root element;

creating a child element into the task root element using the task root element and a task template child element for each child element of the task template root element;

obtaining a list of text strings from the template string specified as a value of a text child node for each text child node in the task template root element; and

creating text child nodes using the task root element and the obtained list of text strings for each text child node in the task template root element.

4. The method according to claim 3 wherein creating a child element into the task root element using a task template root element comprises:

if the task template root element selection is ‘

true’

,

creating a child element into the task root element with a name specified in the task template root element; and

generating attributes and child nodes of the task using the task child element and the task template root element; and

if the task template root element selection is ‘

children’

,

creating child elements and text nodes using the task root element and the task template root element.

5. The method according to claim 1 wherein obtaining a list of text strings from a template string comprises:

initializing a list containing an empty string;

looping through all characters of the template string;

identifying all variables specified in the template string;

updating the list using a variable for each identified variable; and

appending each normal or escaped character in the template string to each string in the list.

6. The method according to claim 5 wherein updating the list using a variable comprises:

looping through all strings in the list;

if the template variable is a scalar variable,

appending the value of the scalar variable to each string in the list; and

if the template variable is a collection variable,

removing a string from the list for each string in the list; and

adding a new string into the list where the new string is obtained by appending an item from the collection value of the variable for each item from the variable value.

7. The method according to claim 1 wherein creating attributes using the task root element, an attribute name and a list of text strings comprises:

if the list contains only a single string,

creating an attribute for the task root element with a given name and the single string as the value; and

if the list contains multiple strings, said method includes;

removing the task root element from its parent element;

looping through all strings in the list; and

creating a new child into the parent task element where the new child is the same as the removed task root element plus a new attribute with a given name and a string as its value for each string in the list.

8. The method according to claim 3 wherein creating text child nodes using the task root element and a list of text strings comprises:

if the list contains only a single string,

creating a text child node into the task root element with the single string as the value; and

if the list contains multiple strings, said method includes;

removing the task root element from its parent element;

looping through all strings in the list; and

creating a new child into the parent task element where the new child is the same as the removed task root element plus a new child text node with a string as its value for each string in the list.

9. The method according to claim 1 wherein creating example tasks comprises providing a standard user interface, the standard user interface providing a capability for a user to specify a period of time of interest, select a subset of event types, and select a type of report to be generated.

10. The method according to claim 1 wherein said received task template comprises list of variables where each variable includes:

a name comprising a string that is unique for all variables within the list;

a description comprising a free text field that describes semantics associated with the variable;

a data type for the variable;

a default value utilized if a user does not provide any value for the variable; and

at least one data integrity constraint.

11. The method according to claim 10 wherein said received task template comprises a target XML document that utilizes the variable from the list of variables in the template.

12. The method according to claim 10 wherein the at least one data integrity constraint includes one or more of:

a minimum value inclusively for ordinal data types;

a minimum value exclusively for ordinal data types;

a maximum value inclusively for ordinal data types;

a maximum value exclusively for ordinal data types;

a minimum size for textual data types and collection data types;

a maximum size for textual data types and collection data types;

a matching regular expression pattern for textual data types; and

arbitrary constraints backed up by a Java class.

13. A computer programmed to create data mining task templates for utilization in data mining activities, said computer programmed to:

receive a task template for discovery of common patterns occurring within data mining events;

create example tasks using the task template by interpreting the task template with specific values of variables, wherein interpreting the task template comprises;

creating a root element of a task from a root element associated with the task template when the variables are a collection data type; and

generating attributes and child nodes of the task using the task root element and the task template root element, wherein to generate attributes and child nodes of the task, said computer is further programmed to;

loop through all the attributes of the task template root element;

obtain a list of text strings from a template string specified as a value of an attribute for each selected attribute in the task template root element;

create attributes using the task root element, an attribute name specified by an attribute in the task template root element, and the obtained list of text strings from a same attribute in the task template root element for each selected attribute in the task template root element; and

create child elements and text nodes using the task root element and the task template root element; and

refine the task template using results returned from execution of the example tasks.

14. The computer programmed according to claim 13 wherein to interpret the task template with variable values, said computer further programmed to:

replace the variables with string representations in attribute values and element text contents;

remove attributes and elements if a corresponding select element has a value of ‘

false’

; and

replace elements with their children when the corresponding select element has a value of ‘

children’

.

15. The computer programmed according to claim 13 wherein to create child elements and text nodes, said computer is further programmed to:

loop through all the child nodes of the task template root element;

create a child element into the task root element using the task root element and a task template child element for each child element of the task template root element;

obtain a list of text strings from a template string specified as a value of a text child node for each text child node in the task template root element; and

create text child nodes using the task root element and the obtained list of text strings for each text child node in the task template root element.

16. The computer programmed according to claim 15 wherein to create a child element into the task root element, said computer is programmed to:

create a child element into the task root element with a name specified in the task template root element and generate attributes and child nodes of the task using the task child element and the task template root element, if the task template root element selection is ‘

true’

; and

create child elements and text nodes using the task root element and the task template root element if the task template root element selection is ‘

children’

.

17. The computer programmed according to claim 13 wherein to obtain a list of text strings from a template string, said computer is programmed to:

initialize a list containing an empty string;

loop through all characters of the template string;

identify all variables specified in the template string;

update the list using a variable for each identified variable; and

append each normal or escaped character in the template string to each string in the list.

18. The computer programmed according to claim 15 wherein to obtain a list of text strings from a template string, said computer is programmed to:

initialize a list containing an empty string;

loop through all characters of the template string;

identify all variables specified in the template string;

update the list using a variable for each identified variable; and

append each normal or escaped character in the template string to each string in the list.

19. The computer programmed according to claim 17 wherein to update the list using a variable, said computer is programmed to:

loop through all strings in the list;

if a template variable is a scalar variable,

append the value of the scalar variable to each string in the list; and

if a template variable is a collection variable,

remove a string from the list for each string in the list; and

add a new string into the list where the new string is obtained by appending an item from the collection value of the variable for each item from the variable value.

20. The computer programmed according to claim 13 wherein to create attributes using the task root element, an attribute name and a list of text strings, said computer is programmed to:

create an attribute for the task root element with a given name and the string as the value if the list contains only a single string; and

if the list contains multiple strings, remove the task root element from its parent element, loop through all strings in the list, and create a new child into the parent task element where the new child is the same as the removed task root element plus a new attribute with the given name and a string as its value for each string in the list.

21. The computer programmed according to claim 15 wherein to create text child nodes using the task root element and a list of text strings, said computer is programmed to:

create a text child node into the task root element with the string as the value if the list contains only a single string; and

if the list contains multiple strings, remove the task root element from its parent element, loop through all strings in the list, and create a new child into the parent task element where the new child is the same as the removed task root element plus a new child text node with a string as its value for each string in the list.

22. The computer programmed according to claim 13 wherein to create example tasks, said computer is programmed to provide a standard user interface, the standard user interface providing a capability for a user to specify a period of time of interest, select a subset of event types, and select a type of report to be generated.

23. The computer programmed according to claim 13 wherein to receive a task template, said computer is programmed to receive a list of variables where each variable includes:

a name comprising a string that is unique for all variables within the list;

a description comprising a free text field that describes semantics associated with the variable;

a data type for the variable;

a default value utilized if a user does not provide any value for the variable; and

at least one data integrity constraint.

24. A system for data mining comprising:

a network comprising a plurality of communicatively coupled computers; and

a server communicatively coupled to said network and accessible by said plurality of computers, said server comprising a memory, said memory comprising at least one data mining task template stored therein for discovery of common patterns occurring within input data, said system operable to;

create tasks using the task template by interpreting the task template with specific variable values, wherein interpreting the task template comprises;

creating a root element of a task from a root element associated with the task template when the variables are a collection data type; and

generating attributes and child nodes of the task using the task root element and the task template root element, wherein generating attributes and child nodes of the task using the task root element and the task template root element comprises;

looping through all the attributes of the task template root element;

obtaining a list of text strings from a template string specified as a value of an attribute for each selected attribute in the task template root element;

creating attributes using the task root element, an attribute name specified by an attribute in the task template root element, and the obtained list of text strings from the same attribute in the task template root element for each selected attribute in the task template root element; and

creating child elements and text nodes using the task root element and the task template root element;

execute the created tasks; and

store data mining results from execution of the tasks within said memory.