Employee selection via multiple neural networks

• US 7,472,097 B1
• Filed: 03/20/2006
• Issued: 12/30/2008
• Est. Priority Date: 03/23/2005
• Status: Active Grant

• Current Assignee: Cadient LLC
• Original Assignee: Kronos Talent Management Incorporated (Seahawk Acquisition Corp.)
• Inventors: Scarborough, David J., Chambless, Bjorn, Thissen-Roe, Anne
• Primary Examiner(s): Holmes; Michael B
• Application Number:
  US11/386,067
• Time in Patent Office: 1,016 Days
• Field of Search: None
• US Class Current: 706/26
• CPC Class Codes: G06N 3/045 Combinations of networks
• CPC Class Codes: G06Q 10/06 Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling

Abstract

A plurality of neural networks or other models can be used in employee selection technologies. A hiring recommendation can be based at least on processing performed by a plurality of neural networks. For example, parallel or series processing by neural networks can be performed. A neural network can be coupled to one or more other neural networks. A binary or other n-ary output can be generated by one or more of the neural networks. In a series arrangement, candidates can be processed sequentially in multiple stages, and those surviving the stages are recommended for hire.

Claims

1. A method comprising:

receiving one or more job performance predictive values for a candidate employee;

performing processing with at least a first neural network, wherein the first neural network generates a prediction indicating whether the candidate employee will be terminated for negative reasons, wherein the negative reasons comprise termination for cause;

performing processing with at least a second neural network, wherein the second neural network generates a prediction indicating a predicted length of service of the candidate employee; and

generating a hiring recommendation for the candidate employee based on the one or more job performance predictive values, wherein the hiring recommendation is based at least on the prediction indicating whether the candidate employee will be terminated for negative reasons according to the processing performed by the first neural network and the hiring recommendation is based at least on whether the predicted length of service meets a threshold length of service according to the processing performed by the second neural network.

2. The method of claim 1 wherein:

the hiring recommendation is based at least on an output of the first neural network and an output of the second neural network.

3. The method of claim 1 wherein:

the first neural network is coupled to the second neural network.

4. The method of claim 1 wherein:

at least one of the neural networks accepts at least one of the one or more job performance predictive values as input.

5. The method of claim 1 wherein:

the processing for the first neural network is performed in a first stage; and

the processing for the second neural network is performed in a second stage.

6. The method of claim 5 wherein:

processing in the second stage is performed responsive to determining the candidate employee has survived the first stage.

7. The method of claim 5 wherein:

a positive hiring recommendation is provided for the candidate employee responsive to determining that the candidate employee has survived the stages.

8. The method of claim 1 wherein:

the processing for the first neural network is performed in parallel with the processing for the second neural network.

9. The method of claim 1 wherein:

the processing for the first neural network is performed in series with the processing for the second neural network.

10. The method of claim 1 wherein:

at least one of the neural networks outputs an n-ary value.

11. The method of claim 1 wherein:

the hiring recommendation comprises an n-ary value; and

the n-ary value is based at least on the processing performed by the first neural network and the processing performed by the second neural network.

12. The method of claim 11 wherein:

the n-ary value indicates whether or not to hire the candidate employee.

13. The method of claim 11 wherein:

the n-ary value comprises a binary value determined via voting by at least one of the neural networks.

14. The method of claim 13 wherein:

at least one of the neural networks is operable to exercise veto power over the binary value.

15. The method of claim 1 wherein:

the processing for the second neural network is performed responsive to determining a favorable outcome with the first neural network.

16. The method of claim 15 wherein:

the first neural network indicates the favorable outcome via an output indicator operable to indicate a binary value.

17. The method of claim 1 wherein at least two of the neural networks are members of a neural network consensus group.

18. The method of claim 1 wherein at least one of the neural networks comprises a neuro-fuzzy network.

29. The method of claim 1 further comprising:

performing processing with an additional neural network, wherein the additional neural network generates a prediction indicating predicted sales of the candidate employee;

wherein the hiring recommendation is based on predicted sales of the candidate employee according to the processing performed by the additional neural network.

30. The method of claim 1 further comprising:

performing processing with an additional neural network, wherein the additional neural network generates a prediction indicating a predicted accident rate of the candidate employee;

wherein the hiring recommendation is based on the predicted accident rate of the candidate employee according to the processing performed by the additional neural network.

31. The method of claim 1 further comprising:

performing processing with an additional neural network, wherein the additional neural network generates a prediction indicating predicted promotions of the candidate employee;

wherein the hiring recommendation is based on predicted promotions of the candidate employee according to the processing performed by the additional neural network.

19. One or more computer-readable storage media having computer-executable instructions for performing a method comprising:

receiving one or more job performance predictive values for a candidate employee;

performing processing with at least a first neural network, wherein the first neural network generates a prediction indicating whether the candidate employee will be terminated for negative reasons, wherein the negative reasons comprise termination for cause;

performing processing with at least a second neural network, wherein the second neural network generates a prediction indicating a predicted length of service of the candidate employee; and

generating a hiring recommendation for the candidate employee based on the one or more job performance predictive values, wherein the hiring recommendation is based at least on the prediction indicating whether the candidate employee will be terminated for negative reasons according to the processing performed by the first neural network and the hiring recommendation is based at least on whether the predicted length of service meets a threshold length of service according to the processing performed by the second neural network.

20. A method comprising:

receiving job application data for a candidate employee;

performing processing on the job application data of the candidate employee with at least a first neural network to generate a first output of the first neural networks wherein the first output indicates whether or not the candidate employee is predicted to be involuntarily terminated;

performing processing on the job application data of the candidate employee with at least a second neural network to generate a second output of the second neural network, wherein the second output indicates a predicted length of service for the candidate employee; and

indicating whether or not to select the candidate employee based on one or more job performance predictive values, wherein the indicating is based at least on the first output of the first neural network and the second output of the second neural network.

21. The method of claim 20 wherein the indicating is operable to generate an output indicating “

hire.”

22. A method comprising:

receiving a plurality of job performance predictive values for a candidate employee;

inputting one or more of the job performance predictive values to a first neural network operable to output a binary value indicative of whether the candidate employee is similar to former employees who were involuntarily terminated; and

inputting one or more of the job performance predictive values to a second neural network operable to output a value indicative of predicted length of service; and

generating a binary value indicating whether to hire the candidate employee based at least on whether the first neural network indicates whether the candidate employee is similar to former employees who were involuntarily terminated and whether or not the second neural network indicates that the predicted length of service is at least a threshold value.

23. The method of claim 22 wherein the second neural network processes the one or more of the job performance predictive values only if the first neural network indicates that the candidate employee is not similar to former employees who were involuntarily terminated.

24. One or more computer-readable storage media comprising:

a hiring recommendation generator operable to generate a hiring recommendation, wherein the hiring generator comprises;

a first neural network, wherein the first neural network is configured to predict involuntary termination; and

a second neural network, wherein the second neural network is configured to predict length of service;

wherein the hiring recommendation is based at least on processing performed by the first neural network and the second neural network.

25. The one or more computer readable media of claim 24 wherein:

the hiring recommendation comprises a binary value indicative of whether or not to hire a candidate employee; and

the binary value is based at least on processing performed by the first neural network and processing performed by the second neural network.

26. The one or more computer readable media of claim 24 wherein:

the first neural network is operable to accept input data indicative of a candidate employee and operable to output a first value indicating predicted performance of the candidate employee if the candidate employee were to be hired; and

a second neural network operable to accept input data indicative of the candidate employee and operable to output a second value indicating predicted performance of the candidate employee if the candidate employee were to be hired;

wherein the hiring recommendation is based at least on the output of the first neural network and the output of the second neural network.

27. An apparatus comprising:

means for generating a hiring recommendation for a candidate employee, wherein the means for generating the hiring recommendation comprises;

means for generating a first output indicative of whether the candidate employee is predicted to be involuntarily terminated; and

means for generating a second output indicative of whether the candidate employee is predicted to have a length of service exceeding a threshold;

wherein the hiring recommendation is based at least on the first output and the second output.

28. A method comprising:

receiving one or more job performance predictive values for a candidate employee;

performing processing with at least a first model configured to predict whether the candidate employee will be involuntarily terminated;

performing processing with at least a second model configured to predict length of service for the candidate employee; and

generating a hiring recommendation for the candidate employee based on the one or more job performance predictive values, wherein the hiring recommendation is based at least on the processing performed by the first model and the processing performed by the second model;

wherein the first and second models are selected from the group consisting of;

a neural network,

a regression model,

a probabilistic model,

a Bayesian model,

a classification tree, and

a discriminant function.