Activity based analytics

• US 10,102,220 B2
• Filed: 06/17/2016
• Issued: 10/16/2018
• Est. Priority Date: 09/27/2013
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Porpora, Gregory A., Spann, Janet L.
• Primary Examiner(s): Reyes, Mariela
• Assistant Examiner(s): Harom, Courtney
• Application Number:
  US15/185,503
• Publication Number:
  US 20160292182A1
• Time in Patent Office: 851 Days
• CPC Class Codes: G06F 16/24568 Data stream processing; Continuous queries
• CPC Class Codes: G06F 16/285 Clustering or classification
• CPC Class Codes: G06F 16/29 Geographical information databases
• CPC Class Codes: G06F 16/335 Filtering based on additional data, e.g. user or group profiles filtering in web context G06F16/9535, G06F16/9536
• CPC Class Codes: G06F 16/35 Clustering; Classification
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06Q 50/26 Government or public services business processes related to the transportation industry G06Q50/40

Abstract

An approach for filtering data is presented. A first geo-hash indicating location information of a person based on profile data or a second geo-hash indicating location information of the person based on an inference is determined to have more characters and is selected as an optimal geo-hash specifying a first geospatial tag. Based on correlations between geospatial tags, time/date stamps, and contextual information, a relationship between first and second entity-metadata elements specifying the person and a vehicle, respectively, and between the person and the vehicle is determined. Representations of the first and second entity-metadata elements are displayed within a regular polygon that includes locations indicated by the geospatial tags. Based on hidden Markov and support vector machine models, a frequent pattern growth algorithm, and a Kohonen map, another activity of the person is predicted.

Claims

1. A method of filtering data, the method comprising the steps of:

a computer determining whether a first geo-hash has more characters than a second geo-hash, the first geo-hash being converted from first location information about a person who is specified by a first area of interest, the first location information extracted from first profile data that describes the person and being a first value of a first geospatial tag, and the second geo-hash being converted from second location information about the person, the second location information inferred by employing a model which is trained by historical data and which uses a k-nearest neighbor distance calculator, and the second location information being a second value of the first geospatial tag;

if the first geo-hash has more characters than the second geo-hash, the computer selecting the first geo-hash as an optimal geo-hash that specifies the first geospatial tag or if the second geo-hash has more characters than the first geo-hash, the computer selecting the second geo-hash as the optimal geo-hash that specifies the first geospatial tag;

based in part on the optimal geo-hash determined by whether the first or second geo-hash has more characters, the computer determining a first correlation between (1) the first geospatial tag and a second geospatial tag and (2) a third geospatial tag, wherein the second geospatial tag along with a first time and date stamp, and first contextual information specifying an activity is included in a first portion of a first group of metadata obtained from data extracted from streaming data and from data at rest, the activity being included in a domain of knowledge associated with law enforcement, and wherein the third geospatial tag is included in a second group of metadata obtained from the data extracted from the streaming data and from the data at rest;

the computer determining a second correlation between (1) the first time and date stamp and a second time and data stamp and (2) a third time and date stamp, wherein the second time and date stamp is included in a second portion of the first group of metadata, and wherein the third time and date stamp is included in the second group of metadata;

the computer determining a third correlation between (1) first and second contextual information and (2) third contextual information, wherein the first contextual information is included in the first portion of the first group of metadata, the second contextual information is included in the second portion of the first group of metadata, and the third contextual information is included in the second group of metadata;

based on the first, second, and third correlations, the computer determining a relationship between a first entity-metadata element specifying the person and a second entity-metadata element specifying a vehicle specified by a second area of interest, and between the person and the vehicle;

based on (1) the relationship between the first and second entity-metadata elements and between the person and the vehicle, and (2) the first, second, and third geospatial tags, the computer displaying representations of the first and second entity-metadata elements within a regular polygon that includes locations indicated by the first, second, and third geospatial tags;

the computer employing a hidden Markov model, which tracks the person and the vehicle;

the computer employing a support vector machine model, which classifies the activity;

the computer employing a frequent pattern growth algorithm, which identifies associations between the activity and one or more other persons;

the computer employing a Kohonen map, which determines a previously unknown activity of the person and the vehicle; and

based on the hidden Markov model, the support vector machine model, the frequent pattern growth algorithm, and the Kohonen map, the computer predicting another activity of the person.

2. The method of claim 1, further comprising the steps of:

subsequent to the step of displaying the representations of the first and second entity-metadata elements, the computer extracting additional data from the streaming data and from the data at rest;

based on the additional extracted data, the computer determining a third entity-metadata element specifying a person of interest, a network of persons, or another vehicle, the third entity-metadata element including a geo-spatial tag that indicates a location within a zone specified by geographic coordinates of a center point and a distance from the center point, the regular polygon being based on the geographic coordinates of the center point and the distance from the center point;

based on the geo-spatial tag included in the third entity-metadata element, the computer generating and displaying a representation of the third entity-metadata element within the regular polygon, which includes the representations of the first and second entity-metadata elements; and

the computer tracking changes to the zone which includes an addition of the third entity-metadata element to the zone.

3. The method of claim 2, wherein the step of determining the third entity-metadata element includes:

performing a continuous query against the extracted additional data using a master key to search for metadata that is geo-referenced based on the center point and the distance from the center point;

based on the master key, cross correlating and disambiguating the metadata that is geo-referenced into a set of groups including at least a first group associated with a person of interest, a second group associated with a network of persons, and a third group associated with another vehicle; and

designating the third entity-metadata element as including a portion of the metadata included in the first, second, or third group in the set of groups.

4. The method of claim 2, further comprising the steps of:

the computer selecting another person, another vehicle, or a network of persons;

the computer extracting other data from the streaming data and from the data at rest;

the computer determining other metadata of the extracted other data;

the computer filtering the extracted other data into another entity-metadata element in another zone based on the other metadata being within metadata boundaries of the other entity-metadata element;

the computer employing complex entity correlation which determines a relationship between the zone and the other zone based on a relationship between the first entity-metadata element and the other entity-metadata element; and

the computer generating and sending an alert about the zone and the other zone being related.

5. The method of claim 1, further comprising the step of:

providing at least one support service for at least one of creating, integrating, hosting, maintaining, and deploying computer-readable program code in the computer, the program code being executed by a processor of the computer to implement the steps of determining whether the first geo-hash has more characters than the second geo-hash, selecting the first geo-hash as the optimal geo-hash if the first geo-hash has more characters than the second geo-hash or selecting the second geo-hash as the optimal geo-hash if the second geo-hash has more characters than the first geo-hash, determining the first correlation, determining the second correlation, determining the third correlation, determining the relationship between the first and second entity-metadata elements and between the person and the vehicle, displaying the representations of the first and second entity-metadata elements, employing the hidden Markov model, employing the support vector machine model, employing the frequent pattern growth algorithm, employing the Kohonen map, and predicting the other activity of the person.

6. A computer system comprising:

a central processing unit (CPU);

a memory coupled to the CPU;

a computer-readable, tangible storage device coupled to the CPU, the storage device containing instructions that are executed by the CPU via the memory to implement a method of filtering data, the method comprising the steps of;

the computer system determining whether a first geo-hash has more characters than a second geo-hash, the first geo-hash being converted from first location information about a person who is specified by a first area of interest, the first location information extracted from first profile data that describes the person and being a first value of a first geospatial tag, and the second geo-hash being converted from second location information about the person, the second location information inferred by employing a model which is trained by historical data and which uses a k-nearest neighbor distance calculator, and the second location information being a second value of the first geospatial tag;

if the first geo-hash has more characters than the second geo-hash, the computer system selecting the first geo-hash as an optimal geo-hash that specifies the first geospatial tag or if the second geo-hash has more characters than the first geo-hash, the computer system selecting the second geo-hash as the optimal geo-hash that specifies the first geospatial tag;

based in part on the optimal geo-hash determined by whether the first or second geo-hash has more characters, the computer system determining a first correlation between (1) the first geospatial tag and a second geospatial tag and (2) a third geospatial tag, wherein the second geospatial tag along with a first time and date stamp, and first contextual information specifying an activity is included in a first portion of a first group of metadata obtained from data extracted from streaming data and from data at rest, the activity being included in a domain of knowledge associated with law enforcement, and wherein the third geospatial tag is included in a second group of metadata obtained from the data extracted from the streaming data and from the data at rest;

the computer system determining a second correlation between (1) the first time and date stamp and a second time and data stamp and (2) a third time and date stamp, wherein the second time and date stamp is included in a second portion of the first group of metadata, and wherein the third time and date stamp is included in the second group of metadata;

the computer system determining a third correlation between (1) first and second contextual information and (2) third contextual information, wherein the first contextual information is included in the first portion of the first group of metadata, the second contextual information is included in the second portion of the first group of metadata, and the third contextual information is included in the second group of metadata;

based on the first, second, and third correlations, the computer system determining a relationship between a first entity-metadata element specifying the person and a second entity-metadata element specifying a vehicle specified by a second area of interest, and between the person and the vehicle;

based on (1) the relationship between the first and second entity-metadata elements and between the person and the vehicle, and (2) the first, second, and third geospatial tags, the computer system displaying representations of the first and second entity-metadata elements within a regular polygon that includes locations indicated by the first, second, and third geospatial tags;

the computer system employing a hidden Markov model, which tracks the person and the vehicle;

the computer system employing a support vector machine model, which classifies the activity;

the computer system employing a frequent pattern growth algorithm, which identifies associations between the activity and one or more other persons;

the computer system employing a Kohonen map, which determines a previously unknown activity of the person and the vehicle; and

based on the hidden Markov model, the support vector machine model, the frequent pattern growth algorithm, and the Kohonen map, the computer system predicting another activity of the person.

7. The computer system of claim 6, wherein the method further comprises the steps of:

subsequent to the step of displaying the representations of the first and second entity-metadata elements, the computer system extracting additional data from the streaming data and from the data at rest;

based on the additional extracted data, the computer system determining a third entity-metadata element specifying a person of interest, a network of persons, or another vehicle, the third entity-metadata element including a geo-spatial tag that indicates a location within a zone specified by geographic coordinates of a center point and a distance from the center point, the regular polygon being based on the geographic coordinates of the center point and the distance from the center point;

based on the geo-spatial tag included in the third entity-metadata element, the computer system generating and displaying a representation of the third entity-metadata element within the regular polygon, which includes the representations of the first and second entity-metadata elements; and

the computer system tracking changes to the zone which includes an addition of the third entity-metadata element to the zone.

8. The computer system of claim 7, wherein the step of determining the third entity-metadata element includes:

performing a continuous query against the extracted additional data using a master key to search for metadata that is geo-referenced based on the center point and the distance from the center point;

based on the master key, cross correlating and disambiguating the metadata that is geo-referenced into a set of groups including at least a first group associated with a person of interest, a second group associated with a network of persons, and a third group associated with another vehicle; and

designating the third entity-metadata element as including a portion of the metadata included in the first, second, or third group in the set of groups.

9. The computer system of claim 7, wherein the method further comprises the steps of:

the computer system selecting another person, another vehicle, or a network of persons;

the computer system extracting other data from the streaming data and from the data at rest;

the computer system determining other metadata of the extracted other data;

the computer system filtering the extracted other data into another entity-metadata element in another zone based on the other metadata being within metadata boundaries of the other entity-metadata element;

the computer system employing complex entity correlation which determines a relationship between the zone and the other zone based on a relationship between the first entity-metadata element and the other entity-metadata element; and

the computer system generating and sending an alert about the zone and the other zone being related.

10. A computer program product, comprising:

a computer-readable, tangible storage device; and

a computer-readable program code stored in the computer-readable, tangible storage device, the computer-readable program code containing instructions that are executed by a central processing unit (CPU) of a computer system to implement a method of filtering data, the method comprising the steps of;

the computer system determining whether a first geo-hash has more characters than a second geo-hash, the first geo-hash being converted from first location information about a person who is specified by a first area of interest, the first location information extracted from first profile data that describes the person and being a first value of a first geospatial tag, and the second geo-hash being converted from second location information about the person, the second location information inferred by employing a model which is trained by historical data and which uses a k-nearest neighbor distance calculator, and the second location information being a second value of the first geospatial tag;

if the first geo-hash has more characters than the second geo-hash, the computer system selecting the first geo-hash as an optimal geo-hash that specifies the first geospatial tag or if the second geo-hash has more characters than the first geo-hash, the computer system selecting the second geo-hash as the optimal geo-hash that specifies the first geospatial tag;

based in part on the optimal geo-hash determined by whether the first or second geo-hash has more characters, the computer system determining a first correlation between (1) the first geospatial tag and a second geospatial tag and (2) a third geospatial tag, wherein the second geospatial tag along with a first time and date stamp, and first contextual information specifying an activity is included in a first portion of a first group of metadata obtained from data extracted from streaming data and from data at rest, the activity being included in a domain of knowledge associated with law enforcement, and wherein the third geospatial tag is included in a second group of metadata obtained from the data extracted from the streaming data and from the data at rest;

the computer system determining a second correlation between (1) the first time and date stamp and a second time and data stamp and (2) a third time and date stamp, wherein the second time and date stamp is included in a second portion of the first group of metadata, and wherein the third time and date stamp is included in the second group of metadata;

the computer system determining a third correlation between (1) first and second contextual information and (2) third contextual information, wherein the first contextual information is included in the first portion of the first group of metadata, the second contextual information is included in the second portion of the first group of metadata, and the third contextual information is included in the second group of metadata;

based on the first, second, and third correlations, the computer system determining a relationship between a first entity-metadata element specifying the person and a second entity-metadata element specifying a vehicle specified by a second area of interest, and between the person and the vehicle;

based on (1) the relationship between the first and second entity-metadata elements and between the person and the vehicle, and (2) the first, second, and third geospatial tags, the computer system displaying representations of the first and second entity-metadata elements within a regular polygon that includes locations indicated by the first, second, and third geospatial tags;

the computer system employing a hidden Markov model, which tracks the person and the vehicle;

the computer system employing a support vector machine model, which classifies the activity;

the computer system employing a frequent pattern growth algorithm, which identifies associations between the activity and one or more other persons;

the computer system employing a Kohonen map, which determines a previously unknown activity of the person and the vehicle; and

based on the hidden Markov model, the support vector machine model, the frequent pattern growth algorithm, and the Kohonen map, the computer system predicting another activity of the person.

11. The program product of claim 10, wherein the method further comprises the steps of:

subsequent to the step of displaying the representations of the first and second entity-metadata elements, the computer system extracting additional data from the streaming data and from the data at rest;

based on the additional extracted data, the computer system determining a third entity-metadata element specifying a person of interest, a network of persons, or another vehicle, the third entity-metadata element including a geo-spatial tag that indicates a location within a zone specified by geographic coordinates of a center point and a distance from the center point, the regular polygon being based on the geographic coordinates of the center point and the distance from the center point;

based on the geo-spatial tag included in the third entity-metadata element, the computer system generating and displaying a representation of the third entity-metadata element within the regular polygon, which includes the representations of the first and second entity-metadata elements; and

the computer system tracking changes to the zone which includes an addition of the third entity-metadata element to the zone.

12. The program product of claim 11, wherein the step of determining the third entity-metadata element includes:

performing a continuous query against the extracted additional data using a master key to search for metadata that is geo-referenced based on the center point and the distance from the center point;

based on the master key, cross correlating and disambiguating the metadata that is geo-referenced into a set of groups including at least a first group associated with a person of interest, a second group associated with a network of persons, and a third group associated with another vehicle; and

designating the third entity-metadata element as including a portion of the metadata included in the first, second, or third group in the set of groups.

13. The program product of claim 11, wherein the method further comprises the steps of:

the computer system selecting another person, another vehicle, or a network of persons;

the computer system extracting other data from the streaming data and from the data at rest;

the computer system determining other metadata of the extracted other data;

the computer system filtering the extracted other data into another entity-metadata element in another zone based on the other metadata being within metadata boundaries of the other entity-metadata element;

the computer system employing complex entity correlation which determines a relationship between the zone and the other zone based on a relationship between the first entity-metadata element and the other entity-metadata element; and

the computer system generating and sending an alert about the zone and the other zone being related.