Data mining and visualization techniques
First Claim
Patent Images
1. A method, comprising:
- processing a dataset with a computer system to determine a number items;
establishing several rules with the computer system, the rules each corresponding to a different association between two or more of the items;
providing a visualization of the rules;
displaying a rule-to-item relationship in the visualization for each one of the rules; and
wherein the rules each correspond to a different location along a first axis of the visualization and the items each correspond to a different location along a second axis of the visualization, and further comprising displaying a first metadata attribute for each of the rules along a third axis of the visualization with the computer system.
2 Assignments
0 Petitions
Accused Products
Abstract
Disclosed are association rule identification and visualization methods, systems, and apparatus. An association rule in data mining is an implication of the form X→Y where X is a set of antecedent items and Y is the consequent item. A unique visualization technique that provides multiple antecedent, consequent, confidence, and support information is disclosed to facilitate better presentation of large quantities of complex association rules.
116 Citations
29 Claims
-
1. A method, comprising:
-
processing a dataset with a computer system to determine a number items;
establishing several rules with the computer system, the rules each corresponding to a different association between two or more of the items;
providing a visualization of the rules;
displaying a rule-to-item relationship in the visualization for each one of the rules; and
wherein the rules each correspond to a different location along a first axis of the visualization and the items each correspond to a different location along a second axis of the visualization, and further comprising displaying a first metadata attribute for each of the rules along a third axis of the visualization with the computer system. - View Dependent Claims (2, 3, 4, 5)
-
-
6. A method, comprising:
-
processing a dataset with a computer to determine several association rules;
providing a visualization of the association rules, the association rules each corresponding to a different one of a number of portions of the visualization;
including a set of identifiers in the visualization for each respective one of the association rules, the identifiers each having a different location along the different one of the portions, one of the identifiers representing a consequent item of the respective one of the association rules, one or more other of the identifiers correspondingly representing one or more antecedent items of the respective one of the association rules; and
wherein the visualization is in a three-dimensional perspective form with the association rules corresponding to a first axis, the set of items corresponding to a second axis, and one or more metadata attributes for each of the association rules corresponding to a third axis. - View Dependent Claims (7, 8, 9, 10, 11, 12)
determining the one or more metadata attributes for each of the association rules with the computer.
-
-
8. The method of claim 7, wherein the one or more metadata attributes number at least two and include a support attribute and a confidence attribute.
-
9. The method of claim 6, wherein the number of portions belong to a planar region of the visualization and the identifiers for the respective one of the association rules each project from the different one of the portions.
-
10. The method of claim 6, wherein said processing includes establishing a set of items from the dataset, the consequent item and the one or more antecedent items for each one of the association rules belonging to the set of items, and further comprising:
-
evaluating the visualization;
selectively screening the set of items; and
generating another visualization after said screening.
-
-
11. The method of claim 6, further comprising:
-
accessing the dataset over a computer network coupled to the computer; and
transmitting an output corresponding to the visualization over the computer network.
-
-
12. The method of claim 6, further comprising modifying the visualization with an operator input device coupled to the computer.
-
13. A computer apparatus, comprising:
- logic to generate a visualization of several association rules, the logic including;
an extraction engine operable to determine a number of items from a dataset;
an association rule mining engine operable to establish the association rules from the items, the rules each corresponding to a different associative relationship between two or more of the items; and
a visualization output generator to define the visualization relative to a first axis, a second axis, and a third axis, the association rules each corresponding to a different location along the first axis, the items each corresponding to a different location along the second axis to provide a rule-to-item relationship for each one of the association rules, and a metadata attribute for each of the association rules being represented along the third axis. - View Dependent Claims (14, 15, 16, 17, 18, 19)
- logic to generate a visualization of several association rules, the logic including;
-
20. A method, comprising:
-
processing a dataset with a computer to determine a number of items;
determining several association rules relative to the items, the association rules including a consequent belonging to the items and one or more antecedents belonging to the items; and
displaying a visualization of the association rules in which the association rules each correspond to a unique position along a first axis, the items each correspond to a unique position along a second axis, and one or more metadata attributes for each of the association rules is represented along a third axis. - View Dependent Claims (21, 22, 23, 24, 25, 26, 27, 28, 29)
determining the one or more metadata attributes for each of the association rules with the computer.
-
-
22. The method of claim 21, wherein the one or more metadata attributes number at least two and include a support attribute and a confidence attribute.
-
23. The method of claim 20, which includes:
- for each of the association rules, representing the consequent and the one or more antecedents in the visualization by a corresponding number of uniquely positioned identifiers each projecting from a planar region of the visualization, the planar region being coincident with the first axis and the second axis.
-
24. The method of claim 20, which includes for each one of the association rules, representing the consequent by a color different than the one or more antecedents in the visualization.
-
25. The method of claim 20, further comprising:
-
accessing the dataset over a computer network coupled to the computer; and
transmitting an output corresponding to the visualization over the computer network.
-
-
26. The method of claim 20, further comprising modifying the visualization with an operator input device coupled to the computer.
-
27. The method of claim 20, which includes:
- for each corresponding one of the association rules, representing the consequent and the one or more antecedents by a corresponding number of identifiers in different positions along the second axis and in the same position along the first axis for the corresponding one of the association rules.
-
28. The method of claim 27, which includes illustrating at least two metadata attributes in the visualization for each of the association rules along a third axis.
-
29. The method of claim 28, which includes for each of the association rules, representing the consequent by a color different than the one or more antecedents in the visualization.
Specification