Analyzing and detecting anomalies in data records using artificial intelligence
First Claim
Patent Images
1. A method comprising:
- creating one or more context insight records representing an indication of a behavior of a service based on raw footprint data related to use of said service by a population of users;
selecting at least one anomaly criteria dimension;
defining possible values for multiple dimensions of said one or more context insight records in response to a result of a scanning of said one or more context insight records;
analyzing, by an analyzer that is a hardware component, said one or more context insight records, in response to said at least one anomaly criteria dimension and to at least one of said possible values defined;
detecting based on said analyzing a dynamically defined anomaly of said service by generating a first generation of candidate anomalies by segmenting said one or more context insight records along a segmentation dimension of said raw footprint data;
generating a consecutive generation of candidate anomalies by modifying, adding or deleting one or more candidate anomalies of said first generation;
said consecutive generation of candidate anomalies further comprising;
modifying said first generation into said consecutive generation by activating a dimension of said one or more context insight records;
selecting, by the hardware analyzing, at least one of each types of dimensions of said context insight records said dimensions comprising;
criteria, segmenting, and testing dimensions;
wherein said criteria dimensions represent characteristics of at least one context insight record population;
wherein said segmenting dimensions divide said at least one context insight record population into several sub-populations;
wherein said testing dimensions comprise criteria by which different sub-populations of the segmenting dimensions are tested;
wherein the generating of at least one of the generations of candidate anomalies is response to, said criteria, segmenting and said testing dimensions selected by the hardware analyzer;
repeating generating consecutive generations of candidate anomalies until a convergence condition holds true; and
assisting in resolving a problem in the service in response to the dynamically defined anomaly.
5 Assignments
0 Petitions
Accused Products
Abstract
A device, system and method for data monitoring, collection and analysis. A method may include creating one or more context insight records representing an indication of a behavior of a service based on raw footprint data related to use of said service by a population of users, and analyzing said one or more context insight records to detect a dynamically defined anomaly of said service.
-
Citations
39 Claims
-
1. A method comprising:
-
creating one or more context insight records representing an indication of a behavior of a service based on raw footprint data related to use of said service by a population of users; selecting at least one anomaly criteria dimension; defining possible values for multiple dimensions of said one or more context insight records in response to a result of a scanning of said one or more context insight records; analyzing, by an analyzer that is a hardware component, said one or more context insight records, in response to said at least one anomaly criteria dimension and to at least one of said possible values defined; detecting based on said analyzing a dynamically defined anomaly of said service by generating a first generation of candidate anomalies by segmenting said one or more context insight records along a segmentation dimension of said raw footprint data; generating a consecutive generation of candidate anomalies by modifying, adding or deleting one or more candidate anomalies of said first generation; said consecutive generation of candidate anomalies further comprising; modifying said first generation into said consecutive generation by activating a dimension of said one or more context insight records; selecting, by the hardware analyzing, at least one of each types of dimensions of said context insight records said dimensions comprising;
criteria, segmenting, and testing dimensions;
wherein said criteria dimensions represent characteristics of at least one context insight record population;
wherein said segmenting dimensions divide said at least one context insight record population into several sub-populations;
wherein said testing dimensions comprise criteria by which different sub-populations of the segmenting dimensions are tested;wherein the generating of at least one of the generations of candidate anomalies is response to, said criteria, segmenting and said testing dimensions selected by the hardware analyzer; repeating generating consecutive generations of candidate anomalies until a convergence condition holds true; and assisting in resolving a problem in the service in response to the dynamically defined anomaly. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17)
-
-
18. An apparatus comprising:
-
an analyzer, that is a hardware component and is configured to create one or more context insight records representing an indication of a behavior of a service based on raw footprint data related to use of said service by a population of users, to select at least one anomaly criteria dimension;
to define possible values for multiple dimensions in response to a result of a scanning of the one or more context insight records; and
to analyze, in response to the at least one anomaly criteria dimension and to at least one of the possible values defined, said one or more context insight records to detect a dynamically defined anomaly of said service;
wherein during an analysis of said one or more context insight the analyzer is configured to;generate a first generation of candidate anomalies by segmenting said one or more context insight records along a segmentation dimension of said raw footprint data; generate a consecutive generation of candidate anomalies by modifying, adding or deleting one or more candidate anomalies of said first generation; and repeat generating consecutive generations of candidate anomalies until a convergence condition holds true, wherein the system is configured to assist in resolving a problem in the service in response to the dynamically defined anomaly; wherein the analyzer is configured to activate a dimension of said one or more context insight records to modify said first generation into said consecutive generation; wherein the analyzer is configured to generate at least one of the generations of candidate anomalies in response to multiple types of dimensions, that comprise criteria dimensions, segmenting dimensions, and testing dimensions, wherein the criteria dimensions represent characteristics of at least one context insight record population;
wherein the segmenting dimensions divide the at least one context insight record population into several sub-populations;
wherein the testing dimensions comprise criteria by which the different sub-populations of the segmenting dimensions may be tested;
wherein at least one dimension of the criteria dimensions, the segmenting dimensions, and the testing dimensions is a dimension selected by the analyzer. - View Dependent Claims (19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35)
-
-
36. A machine-readable medium having stored thereon a set of instructions that, if when executed by a machine, cause the machine to perform a method comprising:
-
creating one or more context insight records representing an indication of a behavior of a service based on raw footprint data related to use of said service by a population of users; selecting at least one anomaly criteria dimension; defining possible values for multiple dimensions in response to a result of a scanning of the one or more context insight records; analyzing said one or more context insight records, in response to said at least one anomaly criteria dimension and to at least one of said possible values defined; detecting based on said analyzing a dynamically defined anomaly of said service by generating a first generation of candidate anomalies by segmenting said one or more context insight records along a segmentation dimension of said raw footprint data; generating a consecutive generation of candidate anomalies by modifying, adding or deleting one or more candidate anomalies of said first generation; wherein generating the consecutive generation comprises; activating a dimension of said one or more context insight records to modify said first generation into said consecutive generation; wherein the generating of at least one of the generations of candidate anomalies is response to multiple types of dimensions, that comprise criteria dimensions, segmenting dimensions, and testing dimensions;
wherein the criteria dimensions represent characteristics of at least one context insight record population;
wherein the segmenting dimensions divide the at least one context insight record population into several sub-populations;
wherein the testing dimensions comprise criteria by which the different sub-populations of the segmenting dimensions may be tested;
wherein at least one dimension of the criteria dimensions, the segmenting dimensions, and the testing dimensions is a dimension selected by the hardware analyzer;repeating generating consecutive generations of candidate anomalies until a convergence condition holds true; and assisting in resolving a problem in the service in response to the dynamically defined anomaly. - View Dependent Claims (37, 38, 39)
-
Specification