Predicting employee performance metrics
Embodiments for predicting employee performance metrics in a computing environment by a processor. An employee performance is predicted over a selected period of time using a dynamic probabilistic model based on historical data, a time window, current and historical observations of the one or more performance metrics, or a combination thereof.
1. A method for predicting employee performance metrics in a computing environment by a processor, comprising:
receiving, by the processor, current and historical observations of one or more performance metrics related to employee performance from a database storing the current and historical observations;
generating, by the processor, a dynamic probabilistic model of a dynamic Bayesian network of a predicted employee performance, the dynamic probabilistic model having a plurality of time steps representative of a time window according to historical data and the current and historical observations of the one or more performance metrics received from the database;
performing, by the processor, a structure learning operation using the generated dynamic probabilistic model, the structure learning operation constrained by a plurality of variables of the one or more performance metrics to generate one or more probabilistic inferences of the plurality of variables; wherein the structure learning operation converts one or more interventional queries of a search operation into one or more probabilistic queries used to generate the one or more probabilistic inferences;
responsive to performing the structure learning operation constrained by the plurality of variables, determining, by the processor, a marginal posterior distribution of a second set of the one or more performance metrics that influence a first set of the one or more performance metrics according to the one or more probabilistic inferences of the plurality of variables; and
outputting, by the processor via a display, a resultant analysis of the predicted employee performance over a selected period of time according to the marginal posterior distribution of the one or more performance metrics of the dynamic probabilistic model; wherein the output further includes displaying certain variables of the plurality of variables of the one or more performance metrics having a greatest amount of influence, at a particular one of the plurality of time steps, on others of the plurality of variables of the one or more performance metrics as relatively compared to all of the plurality of variables of the one or more performance metrics.
2. The method of claim 1 , further including using the dynamic probabilistic model to generate a probability distribution of one or more performance metrics at one or more future time steps.
3. The method of claim 1 , further including initializing a machine learning mechanism using the historical data, the time window, the current and historical observations of the one or more performance metrics, or a combination thereof to learn the dynamic probabilistic model containing dependencies between the one or more performance metrics over the selected period of time.
4. The method of claim 1 , further including learning one or more dependencies between the one or more performance metrics over the selected period of time.
5. The method of claim 1 , further including:
adjusting one or more parameters of one or more performance metrics; and
performing a sensitivity analysis on one or more performance metrics to identify those of the one or more performance metrics having a sensitivity to change based on adjusting the one or more parameters of the one or more performance metrics.
6. The method of claim 1 , further including determining a set of performance metrics having an impact on an alternative set of performance metrics according to a sensitivity analysis.
7. A system for predicting employee performance metrics in a computing environment, comprising:
one or more computers with executable instructions that when executed cause the system to:
receive, by a processor associated with the one or more computers, current and historical observations of one or more performance metrics related to employee performance from a database storing the current and historical observations;
generate, by the processor, a dynamic probabilistic model of a dynamic Bayesian network of a predicted employee performance, the dynamic probabilistic model having a plurality of time steps representative of a time window according to historical data and the current and historical observations of the one or more performance metrics received from the database;
perform, by the processor, a structure learning operation using the generated dynamic probabilistic model, the structure learning operation constrained by a plurality of variables of the one or more performance metrics to generate one or more probabilistic inferences of the plurality of variables; wherein the structure learning operation converts one or more interventional queries of a search operation into one or more probabilistic queries used to generate the one or more probabilistic inferences;
responsive to performing the structure learning operation constrained by the plurality of variables, determine, by the processor, a marginal posterior distribution of a second set of the one or more performance metrics that influence a first set of the one or more performance metrics according to the one or more probabilistic inferences of the plurality of variables; and
output, by the processor via a display, a resultant analysis of the predicted employee performance over a selected period of time according to the marginal posterior distribution of the one or more performance metrics of the dynamic probabilistic model; wherein the output further includes displaying certain variables of the plurality of variables of the one or more performance metrics having a greatest amount of influence, at a particular one of the plurality of time steps, on others of the plurality of variables of the one or more performance metrics as relatively compared to all of the plurality of variables of the one or more performance metrics.
8. The system of claim 7 , wherein the executable instructions use the dynamic probabilistic model to generate a probability distribution of one or more performance metrics at one or more future time steps.
9. The system of claim 7 , wherein the executable instructions initialize a machine learning mechanism using the historical data, the time window, the current and historical observations of the one or more performance metrics, or a combination thereof to learn the dynamic probabilistic model containing dependencies between the one or more performance metrics over the selected period of time.
10. The system of claim 7 , wherein the executable instructions learn one or more dependencies between the one or more performance metrics over the selected period of time.
11. The system of claim 7 , wherein the executable instructions:
adjust one or more parameters of one or more performance metrics; and
perform a sensitivity analysis on one or more performance metrics to identify those of the one or more performance metrics having a sensitivity to change based on adjusting the one or more parameters of the one or more performance metrics.
12. The system of claim 7 , wherein the executable instructions determine a set of performance metrics having an impact on an alternative set of performance metrics according to a sensitivity analysis.
13. A computer program product, by a processor, for predicting employee performance metrics, the computer program product comprising a non-transitory computer-readable storage medium having computer-readable program code portions stored therein, the computer-readable program code portions comprising:
an executable portion that receives, by the processor, current and historical observations of one or more performance metrics related to employee performance from a database storing the current and historical observations;
an executable portion that generates, by the processor, a dynamic probabilistic model of a dynamic Bayesian network of a predicted employee performance, the dynamic probabilistic model having a plurality of time steps representative of a time window according to historical data and the current and historical observations of the one or more performance metrics received from the database;
an executable portion that performs, by the processor, a structure learning operation using the generated dynamic probabilistic model, the structure learning operation constrained by a plurality of variables of the one or more performance metrics to generate one or more probabilistic inferences of the plurality of variables; wherein the structure learning operation converts one or more interventional queries of a search operation into one or more probabilistic queries used to generate the one or more probabilistic inferences;
an executable portion that, responsive to performing the structure learning operation constrained by the plurality of variables, determines, by the processor, a marginal posterior distribution of a second set of the one or more performance metrics that influence a first set of the one or more performance metrics according to the one or more probabilistic inferences of the plurality of variables; and
an executable portion that outputs, by the processor via a display, a resultant analysis of the predicted employee performance over a selected period of time according to the marginal posterior distribution of the one or more performance metrics of the dynamic probabilistic model; wherein the output further includes displaying certain variables of the plurality of variables of the one or more performance metrics having a greatest amount of influence, at a particular one of the plurality of time steps, on others of the plurality of variables of the one or more performance metrics as relatively compared to all of the plurality of variables of the one or more performance metrics.
14. The computer program product of claim 13 , further including an executable portion that uses the dynamic probabilistic model to generate a probability distribution of one or more performance metrics at one or more future time steps.
15. The computer program product of claim 13 , further including an executable portion that:
initializes a machine learning mechanism using the historical data, the time window, the current and historical observations of the one or more performance metrics, or a combination thereof to learn the dynamic probabilistic model containing dependencies between the one or more performance metrics over the selected period of time; and
learns one or more dependencies between the one or more performance metrics over the selected period of time.
16. The computer program product of claim 13 , further including an executable portion that:
adjusts one or more parameters of one or more performance metrics; and
performs a sensitivity analysis on one or more performance metrics to identify those of the one or more performance metrics having a sensitivity to change based on adjusting the one or more parameters of the one or more performance metrics.
17. The computer program product of claim 13 , further including an executable portion that determines a set of performance metrics having an impact on an alternative set of performance metrics according to a sensitivity analysis.