Credit eligibility predictor

1. A computer-implemented method, comprising:

integrating computer-readable program code into a computer system comprising a processor, a computer readable memory in circuit communication with the processor, and a computer readable storage medium in circuit communication with the processor;

the processor executing program code instructions stored on the computer-readable storage medium via the computer readable memory and thereby performing steps of:

extracting, from payroll data of each of plurality of employees of an organization, data that is historically associated to previous instances of certified tax credit eligibility;

normalizing the extracted data with respect to data type and data value;

generating, via a neural network classifier, from an input of the normalized extracted data, multi-class outputs for each employee that indicate strengths of likelihood that each employee is currently eligible for each of a plurality of different tax credits, wherein the neural network classifier is trained on normalized data values and historic success rates for each of a plurality of persons that applied for the tax credits;

filtering the normalized extracted data by removing portions of the normalized extracted data that are associated to ones of the employees that are indicated within the multi-class outputs as having no likelihood that they are currently eligible for any of the plurality of different tax credits, thereby generating a remainder set of normalized extracted data that is associated to the remainder eligible ones of the employees; and

prioritizing an order of application for each of the tax credits for the remainder eligible ones of the employees as a function of respective values and likelihoods of eligibility of the tax credits indicated within the remainder set of normalized extracted data.

2. The method of claim 1 , wherein the normalizing the extracted data comprises:

joining the extracted data with historical information screening information;

labelling the joined extracted data with a data type; and

generating structured data in the labelled data type via a supervised machine learning method from an input of the labelled joined extracted data.

3. The method of claim 2 , wherein the extracted data is at least one of employee name, date of hire, domicile address, business unit address, current and prior employment job titles, job duties, industry types, wage and salary information, military service history, awarded disability status, gender, age, and unique taxpayer identification number.

4. The method of claim 3 , wherein the normalizing the extracted data comprises:

identifying additional data of the employees that is associated to certified tax credit eligibility via analysis of record data that is selected from the group consisting of social and professional networking media, newsfeeds, governmental records, credit report agency records and insurance company records; and

augmenting the extracted data with additional data of the employees.

5. The method of claim 4 , wherein the augmenting comprises identifying via image analysis a military uniform apparel worn by a first of the employees within an image of the first employee that is posted in social media; and

wherein the generating the multi-class output comprises identifying a veteran tax credit associated to the first employee as a function of a nationality of the first employee and the military uniform apparel identified as worn by the first employee.

6. The method of claim 5 , wherein the augmenting comprises identifying, via image analysis and natural language processing, at least one of:

an age of the employee from image data of a posted image or a written statement; and

a disability status of the employee from image data of a posted image; and

wherein the generating the multi-class output comprises identifying at least one of an age credit associated to the first employee as a function of the identified age, and a disability credit associated to the first employee as a function of the identified disability status.

7. The method of claim 1 , further comprising:

training the neural network classifier on a result of a tax credit application executed for one of the prioritized tax credits for one of the remainder eligible ones of the employees.

8. The method of claim 1 , further comprising:

training the neural network classifier as a multi-layer perceptron in a supervised fashion via at least one of backpropagation, genetic algorithms, Lagrange multiplier updates and Bayesian optimization; and

wherein the training of the neural network classifier uses at least one of Keras/Tensorflow, PyTorch, Sklearn, Chainer and Apache MXNet neural network frameworks.

9. The method of claim 8 , wherein the training of the neural network classifier comprises incorporating at least one of L2 regularization, a weight decay regularization and a dropout regularization.

10. The method of claim 1 , wherein the computer-readable program code is provided as a service in a cloud environment.

11. A system, comprising:

a processor;

a computer readable memory in circuit communication with the processor; and

a computer readable storage medium in circuit communication with the processor; and

wherein the processor executes program instructions stored on the computer-readable storage medium via the computer readable memory and thereby:

extracts, from payroll data of each of a plurality of employees of an organization, data that is historically associated to previous instances of certified tax credit eligibility;

normalizes the extracted data with respect to data type and data value;

generates, via a neural network classifier, from an input of the normalized extracted data, multi-class outputs for each employee that indicate strengths of likelihood that each employee is currently eligible for each of a plurality of different tax credits, wherein the neural network classifier is trained on normalized data values and historic success rates for each of a plurality of persons that applied for the tax credits;

filters the normalized extracted data by removing portions of the normalized extracted data that are associated to ones of the employees that are indicated within the multi-class outputs as having no likelihood that they are currently eligible for any of the plurality of different tax credits, thereby generating a remainder set of normalized extracted data that is associated to the remainder eligible ones of the employees; and

prioritizes an order of application for each of the tax credits for the remainder eligible ones of the employees as a function of respective values and likelihoods of eligibility of the tax credits indicated within the remainder set of normalized extracted data.

12. The system of claim 11 , wherein the processor executes the program instructions stored on the computer-readable storage medium via the computer readable memory and thereby normalizes the extracted data by:

joining the extracted data with historical information screening information;

labelling the joined extracted data with a data type; and

generating structured data in the labelled data type via a supervised machine learning method from an input of the labelled joined extracted data.

13. The system of claim 12 , wherein the extracted data is at least one of employee name, date of hire, domicile address, business unit address, current and prior employment job titles, job duties, industry types, wage and salary information, military service history, awarded disability status, gender, age and unique taxpayer identification number.

14. The system of claim 13 , wherein the processor executes the program instructions stored on the computer-readable storage medium via the computer readable memory and thereby normalizes the extracted data by:

identifying additional data of the employees that is associated to certified tax credit eligibility via analysis of record data that is selected from the group consisting of social and professional networking media, newsfeeds, governmental records, credit report agency records and insurance company records; and

augmenting the extracted data with additional data of the employees.

15. The system of claim 14 , wherein the processor executes the program instructions stored on the computer-readable storage medium via the computer readable memory and thereby:

augments the extracted data with additional data of the employees by identifying via image analysis a military uniform apparel worn by a first of the employees within an image of the first employee that is posted in social media; and

generates the multi-class output by identifying a veteran tax credit associated to the first employee as a function of a nationality of the first employee and the military uniform apparel identified as worn by the first employee.

16. The system of claim 15 , wherein the processor executes the program instructions stored on the computer-readable storage medium via the computer readable memory and thereby:

augments the extracted data with additional data of the employees by identifying, via image analysis and natural language processing, at least one of:

an age of the employee from image data of a posted image or of a written statement; and

a disability status of the employee from image data of a posted image; and

generates the multi-class output by identifying at least one of an age credit associated to the first employee as a function of the identified age, and a disability credit associated to the first employee as a function of the identified disability status.

17. The system of claim 16 , wherein the processor executes the program instructions stored on the computer-readable storage medium via the computer readable memory and thereby:

trains the neural network classifier on a result of a tax credit application executed for one of the prioritized tax credits for one of the remainder eligible ones of the employees;

trains the neural network classifier as a multi-layer perceptron in a supervised fashion via at least one of backpropagation, genetic algorithms, Lagrange multiplier updates and Bayesian optimization; and

trains the neural network classifier by using at least one of Keras/Tensorflow, PyTorch, Sklearn, Chainer and Apache MXNet neural network frameworks.

18. A computer program product, comprising:

a computer readable storage medium having computer readable program code embodied therewith, wherein the computer readable storage medium is not a transitory signal per se, the computer readable program code comprising instructions for execution by a processor that cause the processor to:

extract, from payroll data of each of plurality of employees of an organization, data that is historically associated to previous instances of certified tax credit eligibility;

normalize the extracted data with respect to data type and data value;

generate, via a neural network classifier, from an input of the normalized extracted data, multi-class outputs for each employee that indicate strengths of likelihood that each employee is currently eligible for each of a plurality of different tax credits, wherein the neural network classifier is trained on normalized data values and historic success rates for each of a plurality of persons that applied for the tax credits;

filter the normalized extracted data by removing portions of the normalized extracted data that are associated to ones of the employees that are indicated within the multi-class outputs as having no likelihood that they are currently eligible for any of the plurality of different tax credits, thereby generating a remainder set of normalized extracted data that is associated to the remainder eligible ones of the employees; and

prioritize an order of application for each of the tax credits for the remainder eligible ones of the employees as a function of respective values and likelihoods of eligibility of the tax credits indicated within the remainder set of normalized extracted data.

19. The computer program product of claim 18 , wherein the computer readable program code instructions for execution by the processor further cause the processor to normalize the extracted data by:

joining the extracted data with historical information screening information;

labelling the joined extracted data with a data type; and

generating structured data in the labelled data type via a supervised machine learning method from an input of the labelled joined extracted data.