Predictive analytics systems and methods
Various examples of methods and systems are provided for improved predictive analytics. In one example, a method of managing operation of an asset or group of assets of interest includes comparing a generated prediction with one or more prediction range associated with a risk profile assigned to an operational outcome of interest, presenting a notification to an operator in response to the comparison, and incorporating operator-generated input as updated source data for the generation of subsequent predictions. The operator-generated input can comprise an operator-defined selection such as, e.g., acceptance of the notification, or rejection of the notification. The operator-generated input can provide real time or near real time information based upon context-specific knowledge that the operator holds that is substantially independent of historical source data.
1. A computer implemented method of managing operation of at least one asset of interest, comprising the steps:
a. presenting a first generated prediction associated with an operational outcome of interest and sensor data received in real-time, where the first generated prediction includes at least one probability value associated with a likelihood that the operational outcome of interest will occur within a defined future period, where the operational outcome of interest is associated with at least one asset;
b. comparing the first generated prediction with a plurality of prediction thresholds, where each of the plurality of prediction thresholds, independently, comprises an upper and a lower value defining a prediction range associated with a risk profile assigned to the operational outcome of interest, where at least one of the plurality of prediction thresholds comprises a prediction threshold of interest; and
c. presenting a notification to an operator when the comparison indicates that the prediction threshold of interest has been reached or exceeded by the first generated prediction, wherein:
i. the notification provides the operator with information associated with a level of risk that the operational outcome of interest will occur within the defined future period; and
ii. a real-time operator-generated input associated with the notification is receivable from the operator, where the real-time operator-generated input comprises one of at least two real-time operator-defined selections, where the at least two real-time operator-defined selections comprise:
1. a first operator-defined selection comprising operator input associated with an operator decision that the notification is accurate when applied to the current operation of the at least one asset and action on the notification should be scheduled;
2. a second operator selection comprising operator input associated with rejection of the notification, where the second selection results from determination by the operator that, in relation to the current operation of the at least one asset, the notification is incorrect or that the notification is correct, but that the operator will not currently schedule action to address the notification; and
3. a third operator selection comprising operator input associated with placing the notification in a watch queue for a period of time, wherein no additional notifications are provided until there is a change in the prediction threshold of interest; and
iii. information associated with the operator-generated input is configured for use as source data in subsequent predictions generated by the prediction engine; and
d. incorporating at least a portion of the operator-generated input as updated source data in the generation of subsequent predictions from the prediction engine.
2. The computer-implemented method as recited in claim 1 , further including the steps:
providing a historical data management module;
providing a real time data management module;
providing at least one memory that stores computer-executable instructions; and
receiving in real-time sensor data from a plurality of sensor devices; and
providing at least one processor coupled to the historical data management module and real time data management module.
3. The computer-implemented method as recited in claim 2 , wherein the first generated prediction is further obtained from an analysis of historical source data by a prediction engine.
4. The computer-implemented method as recited in claim 2 , further including generating an outcome horizon for the notification, wherein the outcome horizon comprises a period of time that the notification is available for presentation to the operator.
5. The computer-implemented method of claim 4 , wherein the first generated prediction is associated with a prediction threshold indicating that the operational outcome of interest is likely to occur, and if the operational outcome of interest occurs prior to the end of the outcome horizon, the notification is identified as “positive prediction information”.
6. The computer-implemented method of claim 4 , wherein the first generated prediction is associated with a prediction threshold indicating that the operational outcome of interest is likely to occur, and if the operational outcome of interest does not occur prior the end of the outcome horizon, the notification is optionally identified as “negative prediction information”.
7. The computer-implemented method of claim 1 , wherein the operator input of the second operator selection initiates either:
a. suppression of subsequent notifications for a period of time; or
b. re-routing the notification to a second operator for action thereupon.
8. The computer-implemented method of claim 1 , wherein the operator input of the second operator selection comprises a second prediction threshold of interest specified by the operator, and wherein the second operator selection initiates suppression of subsequent notifications until the second prediction threshold of interest is reached or exceeded, wherein when the second prediction threshold of interest is reached or exceeded, the suppressed notification:
a. is made visible to the operator for action thereupon;
b. is suppressed again for a period of time; or
c. re-routed to a second operator for action thereupon.
9. The computer-implemented method of claim 1 , wherein the plurality of prediction thresholds are modifiable in response to an operator input.
10. The computer-implemented method of claim 1 , wherein the operational outcome of interest includes one or more of a fault, a failure, energy use, resource use, production quantity, production quality, compliance with required operational parameters, or an operational cost metric.
11. The computer-implemented method of claim 1 , wherein the at least one asset of interest are associated with a vehicle, industrial machine or industrial process.
12. The computer-implemented method of claim 1 , wherein the historical source data comprises information derived from:
a. one or more sensors associated with an operating system, hardware device, or machine; and
b. recorded information associated with operation of the operating system,
hardware device, or machine.
13. The computer-implemented method of claim 12 , wherein the one or more sensors comprise thermometers, hygrometers, pressure sensors, flow meters, vibration sensors, odometers, accelerometers, ammeters, voltmeters, power meters, digital multi-meters, digital imagers, microscopes, distance measurement devices, tackiness detectors, rigidity detectors, stiffness detectors, alignment detectors, pH detectors, GPS detectors, or combinations thereof.
14. The computer-implemented method of claim 1 , wherein the notification comprises a lead time associated with an action addressing the operational outcome of interest of the notification.
15. A computer system for managing operation of an asset or group of assets of interest, comprising:
a memory configured to store instructions;
a processor disposed in communication with said memory, wherein said processor upon execution of the instructions is configured to:
a. present a first generated prediction associated with an operational outcome of interest and sensor data received in real-time, where the first generated prediction includes at least one probability value associated with a likelihood that the operational outcome of interest will occur within a defined future period, where the operational outcome of interest is associated with at least one asset;
b. compare the first generated prediction with a plurality of prediction thresholds, where each of the plurality of prediction thresholds, independently, comprises an upper and a lower value defining a prediction range associated with a risk profile assigned to the operational outcome of interest, where at least one of the plurality of prediction thresholds comprises a prediction threshold of interest; and
c. present a notification to an operator when the comparison indicates that the prediction threshold of interest has been reached or exceeded by the first generated prediction, wherein:
iii. the notification provides the operator with information associated with a level of risk that the operational outcome of interest will occur within the defined future period; and
iv. a real-time operator-generated input associated with the notification is receivable from the operator, where the real-time operator-generated input comprises one of at least two real-time operator-defined selections, where the at least two real-time operator-defined selections comprise:
1. a first operator-defined selection comprising operator input associated with an operator decision that the notification is accurate when applied to the current operation of the at least one asset and action on the notification should be scheduled;
2. a second operator selection comprising operator input associated with rejection of the notification, where the second selection results from determination by the operator that, in relation to the current operation of the at least one asset, the notification is
incorrect or that the notification is correct, but that the operator will not currently schedule action to address the notification; and
3. a third operator selection comprising operator input associated with placing the notification in a watch queue for a period of time, wherein no additional notifications are provided until there is a change in the prediction threshold of interest; and
iii. information associated with the operator-generated input is configured for use as source data in subsequent predictions generated by the prediction engine; and
d. incorporating at least a portion of the operator-generated input as updated source data in the generation of subsequent predictions from the prediction engine.
16. The computer system as recited in claim 15 , further including:
a historical data management module; and
a real time data management module wherein the processor is coupled to the historical data management module and real time data management module.
17. The computer system as recited in claim 16 , wherein the first generated prediction is further obtained from an analysis of historical source data by a prediction engine.
18. The computer system as recited in claim 17 , wherein the processor is further configured to generate an outcome horizon for the notification, wherein the outcome horizon comprises a period of time that the notification is available for presentation to the operator.
19. The computer system as recited in claim 17 , wherein the first generated prediction is associated with a prediction threshold indicating that the operational outcome of interest is likely to occur, and if the operational outcome of interest occurs prior to the end of the outcome horizon, the notification is identified as “positive prediction information”.
20. The computer system as recited in claim 17 , wherein the first generated prediction is associated with a prediction threshold indicating that the operational outcome of interest is likely to occur, and if the operational outcome of interest does not occur prior the end of the outcome horizon, the notification is optionally identified as “negative prediction information”.