Method and apparatus for efficient and accurate analytics with cross-domain correlation

• US 8,972,484 B2
• Filed: 02/17/2011
• Issued: 03/03/2015
• Est. Priority Date: 02/17/2011
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Naphade, Milind, Sahu, Sambit, Dai, Jing
• Primary Examiner(s): SCOTT, RANDY A
• Application Number:
  US13/029,699
• Publication Number:
  US 20120215829A1
• Time in Patent Office: 1,475 Days
• Field of Search: 709/203, 709/204, 709/205, 340/13.3, 340/440, 340/443
• US Class Current: 709/203
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation

Abstract

A method of generating analytics to provide an analysis of data from distinct data domains includes collecting sensor data from at least two distinct data domains, deriving parameters from the collected data, wherein at least one of the parameters is a first domain parameter derived from one of the data domains and at least another one of the parameters is a second domain parameter derived from the other data domain, providing a data model that enables a user to specify at least one of the first parameters and at least one of the second domain parameters and generate at least one rule based on the selected parameters, and generating analytics that analyze the collected data against the rules to determinate whether the rules have been satisfied and provide results of the analysis to a user of the analytics.

Claims

1. A method of generating analytics to provide an analysis of data from distinct data domains, the method comprising:

collecting data from at least two distinct data domains;

deriving parameters from the collected data comprising first domain parameters derived from a first one of the data domains and second domain parameters derived from a second other one of the data domains;

providing a data model that enables a user to select one of the first domain parameters, one of the second domain parameters, and create a rule based on the selected parameters; and

generating analytics that analyzes the collected data against the rule to determine whether the rule has been satisfied,

wherein the method is performed by a data processing machine,

wherein the analytics predicts whether a single event has occurred or not occurred by using the rule to compare data of the first data domain against the selected first domain parameter and compare data of the second data domain against the selected second domain parameter,

wherein the data of the first data domain is electricity consumption data, the data of the second data domain is water consumption data, the first selected domain parameter is a threshold amount of electricity consumed, and the second selected domain parameter is a threshold amount of water consumed.

2. The method of claim 1, wherein each data domain includes information generated from different information silos, which are incapable of communicating their corresponding information with one another.

3. The method of claim 1, wherein the data is received from Smart Sensors remote from the data processing system.

4. The method of claim 1, wherein at least one of the parameters is a consumption amount of a resource used by a consumer during a consumption period.

5. The method of claim 4, wherein the analytics are configured to determine an average period of non-use of the resource.

6. The method of claim 1, wherein the data includes locations from where the data is sourced and the parameters include the locations.

7. The method of claim 6, wherein the data includes weather data at the locations, and the parameters include weather parameters based on the weather data.

8. The method of claim 1, wherein the analytics provides the prediction to a user.

9. The method of claim 8, wherein the analytics is configured to receive feedback from the user and suppress output of the prediction to the user based on the feedback.

10. The method of claim 1, wherein the single event is use of an electrical appliance.

11. The method of claim 1, wherein the single event is a water leak.

12. The method of claim 1, wherein when the single event has occurred, a water leak has occurred, and when the single event has not occurred, an electrical appliance has consumed water of the water consumption data.

13. A system to provide an analysis of sensor data from distinct data domains to clients, the system comprising:

a computer network;

first sensors configured to measure data from a first data domain and transmit the measured data across the network;

second sensors configured to measure data from a second data domain and transmit the measured data across the network, wherein the data domains are different from one another;

a plurality of client workstations configured to communicate across the network;

a server workstation that is remote from client workstations and configured to communicate with the client workstations and receive sensor data comprising the measured data from the sensors across the network, wherein the server workstation comprises;

a user interface server that enables a user to select one of first domain parameters derived from the first data domain, select one of second domain parameters derived from the second data domain, and generate a correlation rule based on the selected parameters; and

a database server to store the sensor data, the parameters, and the correlation rule; and

analytics that analyze the stored sensor data against the correlation rule to determine whether the rule has been satisfied,

wherein the analytics predicts whether a single event has occurred or not occurred by using the rule to compare data of the first data domain against the selected first domain parameter and compare data of the second data domain against the selected second domain parameter,

wherein the data of the first data domain is water consumption data, the data of the second data domain is electricity consumption data, the first sensors are water meters and the second sensors are electricity meters, the first selected domain parameter is a threshold amount of water consumed, and the second selected domain parameter is a threshold amount of electricity consumed.

14. The system of claim 13, wherein each data domain includes information generated from different information silos, which are incapable of communicating the information with one another.

15. The system of claim 13, wherein the server workstation is a cloud computing infrastructure.

16. The system of claim 13, further comprising an FTP server, wherein the FTP server receives the sensor data from the client workstations, anonymizes the sensor data, and sends the anonymized sensor data to the server workstation, wherein the sensor data includes an identifier that identifies the consumer and the anonymized sensor data excludes the identifier.

17. The system of claim 16, further comprising:

a security server to authenticate clients of the client workstations that desire to connect to the analytics;

a source data staging area to temporarily buffer the anonymized sensor data; and

a data integration server integrating the buffered anonymized sensor data into tables in the database server.

18. A computer program product for generating analytics to provide an analysis of data from distinct data domains, the computer program product comprising:

a non-transitory computer readable storage medium having computer readable program code embodied therewith, the computer readable program, code comprising;

computer readable program code configured to collect data from at least two distinct data domains, derive parameters from the collected data comprising first domain parameters derived from a first one of the data domains and second domain parameters derived from a second other one of the data domains, provide a data model that enables a user to select one of the first domain parameters and one of the second domain parameters, and generate a rule based on the selected parameters, and generate analytics that analyze the collected data against the rule to determine whether the rule has been satisfied,

wherein the analytics predicts whether a single event has occurred or not occurred by using the rule to compare data of the first data domain against the selected first domain parameter and compare data of the second data domain against the selected second domain parameter,

wherein the data of the first data domain is electricity consumption data, the data of the second data domain is water consumption data, the first selected domain parameter is a threshold amount of electricity consumed, and the second selected domain parameter is a threshold amount of water consumed.