Event identification in sensor analytics
First Claim
Patent Images
1. A method of detecting an event anomaly comprising:
- receiving, by a data processing apparatus, one or more individualized data points, wherein each individualized data point represents a spatial or temporal event and has a unique identifier associated with it;
distributing, by the data processing apparatus, the one or more individualized data points across a grid, wherein the grid includes one or more cells;
determining, by the data processing apparatus, an event likelihood ratio for each of one or more of the grid cells, wherein determining the event likelihood ratio for a respective grid cell comprises determining an expected number of individualized data points in the respective grid cell;
identifying, by the data processing apparatus, one or more event clusters, based on the event likelihood ratio, wherein each event cluster includes one or more of the grid cells, thus defining a region on the grid;
generating, by the data processing apparatus, multiple replicas of the region;
performing, by the data processing apparatus, simulations on the multiple replicas to obtain a significance value for each of the one or more event clusters; and
storing, by the data processing apparatus, the one or more event clusters in a data repository, based on a comparison of the significance value to a threshold.
27 Assignments
0 Petitions
Accused Products
Abstract
A method of detecting an event anomaly includes receiving one or more data points, in which each data point represents a spatial or temporal event; associating a unique identifier with each of the one or more data points to obtain one or more individualized data points; distributing the one or more individualized data points across a grid, in which the grid includes one or more cells; determining an event likelihood ratio for one or more of the grid cells; identifying one or more event clusters, in which each event cluster includes one or more of the grid cells; and storing in a data repository an event cluster having a significance level above a threshold significance level.
-
Citations
48 Claims
-
1. A method of detecting an event anomaly comprising:
-
receiving, by a data processing apparatus, one or more individualized data points, wherein each individualized data point represents a spatial or temporal event and has a unique identifier associated with it; distributing, by the data processing apparatus, the one or more individualized data points across a grid, wherein the grid includes one or more cells; determining, by the data processing apparatus, an event likelihood ratio for each of one or more of the grid cells, wherein determining the event likelihood ratio for a respective grid cell comprises determining an expected number of individualized data points in the respective grid cell; identifying, by the data processing apparatus, one or more event clusters, based on the event likelihood ratio, wherein each event cluster includes one or more of the grid cells, thus defining a region on the grid; generating, by the data processing apparatus, multiple replicas of the region; performing, by the data processing apparatus, simulations on the multiple replicas to obtain a significance value for each of the one or more event clusters; and storing, by the data processing apparatus, the one or more event clusters in a data repository, based on a comparison of the significance value to a threshold. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44)
-
-
45. A non-transitory computer readable medium having stored thereon a computer program that, when executed, causes a computer to perform the steps of:
-
receiving one or more individualized data points, wherein each individualized data point represents a spatial or temporal event and has a unique identifier associated with it; distributing the one or more individualized data points across a grid, wherein the grid includes one or more cells; determining an event likelihood ratio for each of one or more of the grid cells, wherein determining the event likelihood ratio for a respective grid cell comprises determining an expected number of individualized data points in the respective grid cell; identifying one or more event clusters, based on the event likelihood ratio, wherein each event cluster includes one or more of the grid cells, thus defining a region on the grid; generating multiple replicas of the region; performing simulations on the multiple replicas to obtain a significance value for each of the one or more event clusters; and storing the one or more event clusters in a data repository, based on a comparison of the significance value to a threshold.
-
-
46. A system for detecting an event anomaly comprising:
-
a network; one or more user devices in communication with the network; and a processor in communication with the network, wherein the processor is for; i) receiving one or more individualized data points representing data originating from the one or more user devices, wherein each individualized data point represents a spatial or temporal event and has a unique identifier associated with it; ii) distributing the one or more individualized data points across a grid, wherein the grid includes one or more cells; iii) determining an event likelihood ratio for each of one or more of the grid cells, wherein determining the event likelihood ratio for a respective grid cell comprises determining an expected number of individualized data points in the respective grid cell; iv) identifying one or more event clusters, based on the event likelihood ratio, wherein each event cluster includes one or more of the grid cells, thus defining a region on the grid; v) generating multiple replicas of the region; vi) performing simulations on the multiple replicas to obtain a significance value for each of the one or more event clusters; and vii) storing the one or more event clusters in a data repository, based on a comparison of the significance value to a specified threshold value.
-
-
47. A system comprising:
-
means for distributing one or more individualized data points across a grid, wherein the grid includes one or more cells; means for detecting an event likelihood ratio for each of the one or more cells of the grid, wherein detecting the event likelihood ratio for a respective grid cell comprises determining an expected number of individualized data points in the one or more cells; means for identifying one or more event cluster based on the event likelihood ratio, wherein each event cluster includes one or more of the grid cells, thus defining a region on the grid; means for generating multiple replicas of the region; means for performing simulations on the multiple replicas to obtain a significance value for each of the one or more event cluster; and means for storing the one or more event clusters in a data repository, based on a comparison of the significance value to a threshold.
-
-
48. A system comprising:
-
means for distributing one or more individualized data points across a grid, wherein the grid includes one or more cells; means for detecting an event likelihood ratio for each of the one or more cells of the grid, wherein detecting the event likelihood ratio for a respective grid cell comprises determining an expected number of individualized data points in the one or more cells; means for identifying one or more event cluster based on the event likelihood ratio, wherein each event cluster includes one or more of the grid cells, thus defining a region on the grid; means for generating multiple replicas of the region; means for performing simulations on the multiple replicas to obtain a significance value for each of the one or more event cluster, wherein the significance value corresponds to a ratio of the number of simulations that result in a maximum significance score that is greater than the event likelihood ratio relative to a total number of simulations performed; and means for storing the one or more event clusters in a data repository when the significance value is determined to be less than a threshold risk level.
-
Specification