System and method using sampling for allocating web page placements in online publishing of content

1. A computer system for publishing content, comprising:

a memory storing computer-executable instructions;

a processor device operably coupled with the memory, said processor device executing a multi-armed bandit engine for learning the valuation of a plurality of content items through sampling by scheduling the content items for web page placements to optimize payments for maximizing revenue;

wherein the multi-armed bandit engine schedules the content items for web page placement by:

selecting an optimal set of content items to sample from the plurality of content items;

allocating a web page placement for a content item in the optimal set of content items to maximize revenue, wherein allocating said web page placement comprises:

sorting a plurality of web page placements in ascending order by normalized click probability;

sorting the optimal set of content items in descending order by payoff;

matching the plurality of web page placements in ascending order with a plurality of content items of the optimal set of content items in descending order; and

allocating the plurality of web page placements matched to the plurality of content items of the optimal set of content items;

sampling the plurality of content items allocated to the plurality of web page placements;

updating a click-through rate for the content item sampled; and

calculating a charge for the content item allocated to the web page placement; and

a storage operably coupled to the processor device for storing a plurality of valuations each associated with a content item allocated to a web page placement for online publishing of content.

2. The system of claim 1 further comprising a model generator for creating a multi-armed bandit model used by the multi-armed bandit engine.

3. The system of claim 1 further comprising a payoff optimizer operably coupled to the multi-armed bandit engine for optimizing payments for the content items sampled online to maximize revenue.

4. A computer-implemented method for publishing content, comprising:

using a processor device configured to perform:

selecting an optimal set of content items to sample from a plurality of content items;

allocating a web page placement for a content item in the optimal set of content items to maximize revenue, wherein allocating said web page placement comprises:

sorting a plurality of web page placements in ascending order by normalized click probability;

sorting the optimal set of content items in descending order by payoff;

matching the plurality of web page placements in ascending order with a plurality of content items of the optimal set of content items in descending order; and

allocating the plurality of web page placements matched to the plurality of content items of the optimal set of content items;

sampling the plurality of content items allocated to the plurality of web page placements;

updating a click-through rate for the content item sampled; and

calculating a charge for the content item allocated to the web page placement.

5. The method of claim 4 further comprising receiving the plurality of content items.

6. The method of claim 4 further comprising initializing a value per click of each content item of the plurality of content items.

7. The method of claim 4 further comprising initializing the click-through rate for each of the plurality of content items.

8. The method of claim 4 further comprising initializing a price charged for a click of a content item for each of the plurality of content items.

9. The method of claim 4 further comprising initializing an exposure parameter for each of a plurality of publishers on a first visit in an online publishing system.

10. The method of claim 4 wherein selecting an optimal set of content items to sample from a plurality of content items comprises:

determining a list of the plurality of content items;

removing content items with a payoff lower than a threshold from the list; and

outputting the remaining list of the plurality of content items as the optimal set of content items to sample.

11. The method of claim 4 wherein allocating a web page placement for a content item in the optimal set of content items to maximize revenue comprises:

sorting a plurality of web page placements in rank order by normalized click probability;

obtaining the highest ranked unallocated web page placement in the rank order;

determining an optimal subset of content items from the optimal set of content items;

randomly selecting the content item from the optimal subset of content items;

allocating the highest ranked unallocated web page placement to the content item randomly selected; and

removing the content item randomly selected from the optimal set of content items.

12. The method of claim 4 further comprising updating a payoff rate for the sampled content item.

13. The method of claim 4 further comprising updating a confidence parameter for the sampled content item.

14. The method of claim 4 wherein updating a click-through rate for the content item sampled comprises updating the click-through rate for the content item sampled using a normalized number of clicks.

15. The method of claim 14 wherein updating the click-through rate for the content item sampled using a normalized number of clicks comprises updating the click-through rate for the content item sampled using probability constants.

16. A non-transitory computer-readable medium having computer-executable instructions for performing;

selecting an optimal set of content items to sample from a plurality of content items;

allocating a web page placement for a content item in the optimal set of content items to maximize revenue, wherein allocating said web page placement comprises:

sorting a plurality of web page placements in ascending order by normalized click probability;

sorting the optimal set of content items in descending order by payoff;

matching the plurality of web page placements in ascending order with a plurality of content items of the optimal set of content items in descending order; and

allocating the plurality of web page placements matched to the plurality of content items of the optimal set of content items;

sampling the plurality of content items allocated to the plurality of web page placements;

updating a click-through rate for the content item sampled; and

calculating a charge for the content item allocated to the web page placement.

17. A computer-implemented method for publishing content, comprising:

using a processor device configured to perform:

selecting an optimal set of content items to sample from a plurality of content items;

allocating a web page placement for a content item in the optimal set of content items to maximize revenue, wherein allocating said web page placement comprises:

sorting a plurality of web page placements in rank order by normalized click probability;

obtaining a highest ranked unallocated web page placement in the rank order;

determining an optimal subset of content items from the optimal set of content items;

randomly selecting a content item from the optimal subset of content items;

allocating the highest ranked unallocated web page placement to the content item randomly selected; and

removing the content item randomly selected from the optimal set of content items;

sampling the content item allocated to the web page placement;

updating a click-through rate for the content item sampled; and

calculating a charge for the content item allocated to the web page placement.

18. The computer-implemented method of claim 17 wherein selecting the optimal set of content items to sample from the plurality of content items comprises:

generating a list of the plurality of content items;

removing content items with a payoff lower than a threshold from the list; and

outputting the list of the plurality of content items as the optimal set of content items to sample.