Providing training information for training a categorizer
First Claim
Patent Images
1. A system, comprising:
- at least one processor;
a data set comprising a plurality of cases;
a search engine executable on the at least one processor to receive a query relating to at least one category and to identify cases within the data set that match the query, the identified cases being unlabeled with respect to the category, wherein the search engine is to identify the cases that match the query without using a categorizer that determines whether or not the cases belong to the category;
a confirmation module executable on the at least one processor to;
receive a first user indication in a user interface that a first case of the identified cases belongs to the category, and a second user indication in the user interface that a second case of the identified cases does not belong to the category,in response to receiving the first and second user indications in the user interface, modify training information for training the categorizer, the confirmation module modifying the training information by adding the first case to a positive training set of cases, and adding the second case to a negative training set of cases; and
a training module executable on the at least one processor to modify the categorizer based on the positive and negative training sets.
9 Assignments
0 Petitions
Accused Products
Abstract
A method and system of providing training information for training a categorizer includes receiving a query relating to at least one category and identifying at least one case within a data set that matches the query. The method and system receives one of a first indication that the identified at least one case belongs to the category, and a second indication that the identified at least one case does not belong to the category. Training information is modified based on receiving one of the first indication and second indication.
-
Citations
44 Claims
-
1. A system, comprising:
-
at least one processor; a data set comprising a plurality of cases; a search engine executable on the at least one processor to receive a query relating to at least one category and to identify cases within the data set that match the query, the identified cases being unlabeled with respect to the category, wherein the search engine is to identify the cases that match the query without using a categorizer that determines whether or not the cases belong to the category; a confirmation module executable on the at least one processor to; receive a first user indication in a user interface that a first case of the identified cases belongs to the category, and a second user indication in the user interface that a second case of the identified cases does not belong to the category, in response to receiving the first and second user indications in the user interface, modify training information for training the categorizer, the confirmation module modifying the training information by adding the first case to a positive training set of cases, and adding the second case to a negative training set of cases; and a training module executable on the at least one processor to modify the categorizer based on the positive and negative training sets. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18)
-
-
19. A method, comprising:
-
receiving, by a system having a processor, a query relating to at least a first category to search cases stored in a data set; identifying, by the system, a first group of cases in the data set matching the query, the first group of cases being unlabeled with respect to the first category, wherein the first group of cases that are unlabeled with respect to the first category have not been labeled by a categorizer for determining whether or not cases belong to the first category; receiving, by the system, a first user indication in a user interface that at least a first case in the first group belongs to the first category, and a second user indication in the user interface that at least a second case in the first group does not belong to the first category; modifying, by the system, training information for training the categorizer in response to receiving the first and second user indications, wherein modifying the training information comprises adding the first case to a positive training set of cases and adding the second case to a negative training set of cases; and modifying, by the system, the categorizer based on the positive and negative training sets. - View Dependent Claims (20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36)
-
-
37. An article comprising at least one non-transitory storage medium containing instructions that when executed cause a computer to:
-
store a data set comprising a plurality of cases not labeled with respect to a category; receive a first query relating to at least the category; identify cases within the data set that match the first query, the identified cases unlabeled with respect to the category, wherein the identified cases that are unlabeled with respect to the category have not been labeled by a categorizer for determining whether or not the cases belong to the category; receive a first user indication in a user interface that a first case of the identified cases belongs to the category, and a second user indication in the user interface that a second case of the identified cases does not belong to the category; modify training information for training the categorizer in response to receiving the first user indication and the second user indication, wherein modifying the training information comprises adding the first case to a positive training set of cases and adding the second case to a negative training set of cases; and modifying the categorizer based on the positive and negative training sets. - View Dependent Claims (38, 39, 40, 41, 42, 43, 44)
-
Specification