Selecting attributes by progressive sampling to generate digital predictive models

1. In a digital medium environment, a computer-implemented method of selecting attributes for generating digital predictive models, comprising:

determining an initial set of attributes corresponding to a repository of data samples, the initial set of attributes having an initial number of attributes;

utilizing the initial set of attributes to identify a coarse sample set having a coarse number of data samples from the repository of data samples;

selecting a focused subset of the initial set of attributes based on the coarse sample set, wherein the focused subset of the initial set of attributes has a focused number of attributes less than the initial number of attributes;

utilizing the focused subset of the initial set of attributes to identify a refined sample set from the repository of data samples having a refined number of data samples larger than the coarse number of data samples; and

selecting a final subset of attributes from the focused subset of the initial set of attributes, wherein the final subset of attributes have a final number of attributes less than the focused number of attributes;

utilizing the final subset of attributes to generate a final sample set having a final number of data samples larger than the refined number of data samples; and

generating a digital predictive model for predicting a target attribute based on the final sample set comprising the final subset of attributes.

2. The computer-implemented method of claim 1 , further comprising:

determining one or more attributes of a user client device; and

generating a predicted target attribute by processing the one or more attributes of the user client device utilizing the digital predictive model.

3. The computer-implemented method of claim 2 , further comprising providing digital content to the user client device based on the predicted target attribute.

4. The computer-implemented method of claim 2 , further comprising generating the predicted target attribute by generating a predicted action of the user client device with regard to a user interface element.

5. The computer-implemented method of claim 2 , wherein determining the one or more attributes of the user client device comprises determining at least one of location, time, language, demographic information, action, device type, or software type corresponding to the user client device.

6. The computer-implemented method of claim 1 , wherein determining the initial set of attributes comprises:

providing, for display via a graphical user interface, a plurality of selectable attributes;

and determining a user selection of the initial set of attributes from the plurality of selectable attributes.

7. The computer-implemented method of claim 1 , wherein generating the digital predictive model for predicting the target attribute based on the final sample set utilizes fewer computing resources than generating the digital predictive model utilizing the repository of data samples and the initial number of attributes.

8. The computer-implemented method of claim 1 , further comprising:

comparing the initial number of attributes to a threshold; and

utilizing the initial set of attributes to identify the coarse sample set in response to determining that the initial number of attributes satisfies the threshold.

9. The computer-implemented method of claim 1 , further comprising determining the coarse number of data samples based on an attribute threshold and the final number of data samples.

10. The computer-implemented method of claim 9 , further comprising determining the attribute threshold and the final number of data samples based on user selection via a computing device.

11. A system for improving digital attribute selection in generating digital predictive models, comprising:

one or more memory devices comprising a repository of data samples, wherein the repository of data samples reflects a plurality of samples and a corresponding set of attributes for each sample;

at least one server device configured to cause the system to:

determine an initial set of attributes corresponding to the repository of data samples, the initial set of attributes having an initial number of attributes;

utilize the initial set of attributes to identify a coarse sample set having a coarse number of data samples from the repository of data samples;

select a focused subset of the initial set of attributes based on the coarse sample set, wherein the focused subset of the initial set of attributes has a focused number of attributes less than the initial number of attributes;

utilize the focused subset of the initial set of attributes to identify a refined sample set from the repository of data samples having a refined number of data samples larger than the coarse number of data samples; and

select a final subset of attributes from the focused subset of the initial set of attributes, wherein the final subset of attributes have a final number of attributes less than the focused number of attributes;

utilize the final subset of attributes to generate a final sample set having a final number of data samples larger than the refined number of data samples; and

generate a digital predictive model for predicting a target attribute based on the final sample set comprising the final subset of attributes.

12. The system of claim 11 , wherein the at least one server device is further configured to:

determine one or more attributes of a user client device;

generate a predicted target attribute by processing the one or more attributes of the user client device utilizing the digital predictive model; and

provide digital content to the user client device based on the predicted target attribute.

13. The system of claim 12 , wherein the at least one server device is further configured to:

determine the initial set of attributes by:

providing, for display via a graphical user interface, a plurality of selectable attributes; and

identifying a user selection of the initial set of attributes from the plurality of selectable attributes; and

generate the predicted target attribute by generating a predicted action of the user client device with regard to a user interface element.

14. The system of claim 11 , further comprising utilizing the initial set of attributes to identify the coarse sample set in response to determining that the initial number of attributes satisfies a threshold.

15. The system of claim 11 , further comprising determining the coarse number of data samples based on the final number of data samples.

16. A non-transitory computer readable medium storing instructions thereon that, when executed by at least one processor, cause a computing device to:

determine an initial set of attributes corresponding to a repository of data samples, the initial set of attributes having an initial number of attributes;

utilize the initial set of attributes to identify a coarse sample set having a coarse number of data samples from the repository of data samples;

select a focused subset of the initial set of attributes based on the coarse sample set, wherein the focused subset of the initial set of attributes has a focused number of attributes less than the initial number of attributes;

utilize the focused subset of the initial set of attributes to identify a refined sample set from the repository of data samples having a refined number of data samples larger than the coarse number of data samples; and

select a final subset of attributes from the focused subset of the initial set of attributes, wherein the final subset of attributes have a final number of attributes less than the focused number of attributes;

utilize the final subset of attributes to generate a final sample set having a final number of data samples larger than the refined number of data samples; and

generate a digital predictive model for predicting a target attribute based on the final sample set comprising the final subset of attributes.

17. The non-transitory computer readable medium of claim 16 , further comprising instructions that, when executed by the at least one processor, cause the computing device to:

determine one or more attributes of a user client device; and

generate a predicted action of the user client device by processing the one or more attributes of the user client device utilizing the digital predictive model.

18. The non-transitory computer readable medium of claim 16 , further comprising instructions that, when executed by the at least one processor, cause the computing device to:

determine that the initial number of attributes satisfies a threshold; and

utilize the initial set of attributes to identify the coarse sample set.

19. The non-transitory computer readable medium of claim 16 , further comprising instructions that, when executed by the at least one processor, cause the computing device to determine the coarse number of data samples proportionate to an attribute threshold and inversely proportionate to the final number of data samples.

20. The non-transitory computer readable medium of claim 19 , further comprising instructions that, when executed by the at least one processor, cause the computing device to determine the attribute threshold and the final number of data samples based on user selection via a user interface of a client device.