Computing a quantification measure associated with cases in a category
First Claim
Patent Images
1. A method comprising:
- identifying, in response to a query relating to at least one category, plural cases from a set of cases;
receiving, by a system having a processor, indications corresponding to user confirmation or user disconfirmation of individual ones of the plural cases as belonging to the category, wherein user confirmation of a first of the plural cases as belonging to the category includes user selection of a first user-selectable element in a display screen, and wherein user disconfirmation of a second of the plural cases as belonging to the category includes user selection of a second user-selectable element in the display screen;
in response to the indications corresponding to the user confirmation of the first case and the user disconfirmation of the second case, adding, by the system, the first case to a positive training set, and the second case to a negative training set;
training, by the system, a categorizer based on the positive and negative training sets;
computing, by the system, a quantification measure associated with cases in the category based on output from the trained categorizer, wherein each of at least some of the cases in the category has a data field, and wherein computing the quantification measure comprises computing an aggregate of the data fields of the at least some of the cases in the category; and
adjusting, by the system, the quantification measure to calibrate for inaccuracy of the trained categorizer.
8 Assignments
0 Petitions
Accused Products
Abstract
At least one case from a set of cases is identified in response to a query relating to at least one category. An indication is received regarding whether the at least one case belongs to the category. A categorizer is trained based on the received indication. A quantification measure associated with cases in the category is computed based on output from the categorizer.
66 Citations
24 Claims
-
1. A method comprising:
-
identifying, in response to a query relating to at least one category, plural cases from a set of cases; receiving, by a system having a processor, indications corresponding to user confirmation or user disconfirmation of individual ones of the plural cases as belonging to the category, wherein user confirmation of a first of the plural cases as belonging to the category includes user selection of a first user-selectable element in a display screen, and wherein user disconfirmation of a second of the plural cases as belonging to the category includes user selection of a second user-selectable element in the display screen; in response to the indications corresponding to the user confirmation of the first case and the user disconfirmation of the second case, adding, by the system, the first case to a positive training set, and the second case to a negative training set; training, by the system, a categorizer based on the positive and negative training sets; computing, by the system, a quantification measure associated with cases in the category based on output from the trained categorizer, wherein each of at least some of the cases in the category has a data field, and wherein computing the quantification measure comprises computing an aggregate of the data fields of the at least some of the cases in the category; and adjusting, by the system, the quantification measure to calibrate for inaccuracy of the trained categorizer. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18)
-
-
19. A system comprising:
-
at least one processor; a data set comprising cases; a search engine to identify plural cases from the data set in response to a query relating to at least one category; a confirmation module executable on the at least one processor to receive indications corresponding to user confirmation or user disconfirmation of individual ones of the plural cases as belonging to the category, wherein user confirmation of a first of the plural cases as belonging to the category includes user selection of a first user-selectable element in a display screen, and wherein user disconfirmation of a second of the plural cases as belonging to the category includes user selection of a second user-selectable element in the display screen, the confirmation module to further, in response to the indications corresponding to the user confirmation of the first case and the user disconfirmation of the second case, add the first case to a positive training set, and the second case to a negative training set; a categorizer to be trained based on the positive training set and negative training set; and a quantifier to compute a quantification measure associated with cases in the category based on output from the trained categorizer, wherein each of the cases in the category has a data field, and the quantification measure is computed by aggregating the data fields of the cases in the category, and wherein the quantifier is to adjust, to calibrate for inaccuracy of the trained categorizer, the quantification measure based on a measure of accuracy of the trained categorizer. - View Dependent Claims (20, 21, 22, 23)
-
-
24. An article comprising at least one non-transitory storage medium containing instructions that when executed cause a system to:
-
identify, in response to a query relating to at least one category, plural cases from a set of cases; receive indications corresponding to user confirmation or user disconfirmation of individual ones of the plural cases as belonging to the category, wherein user confirmation of a first of the plural cases as belonging to the category includes user selection of a first user-selectable element in a display screen, and wherein user disconfirmation of a second of the plural cases as belonging to the category includes user selection of a second user-selectable element in the display screen; in response to the indications corresponding to the user confirmation of the first case and the user disconfirmation of the second case, adding the first case to a positive training set, and the second case to a negative training set; train a categorizer based on the positive and negative training sets; compute a quantification measure associated with cases in the category based on output from the trained categorizer, wherein each of the cases in the category has a data field, and wherein computing the quantification measure comprises computing an aggregate of the data fields of the cases in the category; and adjust, to calibrate for inaccuracy of the trained categorizer, the quantification measure based on a measure of accuracy of the trained categorizer.
-
Specification