Generating keywords by associative context with input words

1. A system, comprising:

one or more processors; and

a machine-readable storage medium storing a set of instructions that; when executed by the one or more processors, cause a machine to perform operations comprising:

accessing, by the one or more processors, a plurality of inventory item listings accessible over a network, the plurality of inventory item listings including one or more text strings;

creating, by the one or more processors, inventory word vectors for at least some of the text strings of the plurality of inventory item listings;

receiving, by the one or more processors, a user input including an input word, the user input being input by a user to access a primary media content;

creating, by the one or more processors, an input word vector for the input word;

calculating, by the one or more processors, cosine similarities between the input word vector and the inventory word vectors; and

analyzing, by the one or more processors, the calculated cosine similarities to determine one or more keywords from the one or more text strings, the one or more keywords being from the text strings having inventory word vectors with the highest cosine similarities to the input word vector.

2. The system of claim 1 ; the operations further comprising:

associating the one or more keywords with a secondary media content; and providing the secondary media content to be presented to the user.

3. The system of claim 2 , the primary media content being a set of search results.

4. The system of claim 2 , the secondary media content being a targeted advertisement.

5. The system of claim 2 , the primary media content and the secondary media content being presented to the user simultaneously.

6. The system of claim 1 , the plurality of inventory item listings being listings from an online marketplace.

7. The system of claim 1 , the one or more text strings from each of the plurality of inventory item listings being from titles of the inventory item listings.

8. The system of claim 1 , the operations further comprising:

creating revised inventory word vectors for at least some of the same text strings for which the inventory word vectors were originally created;

calculating cosine similarities between the input word vector and the revised inventory word vectors; and

analyzing the calculated cosine similarities to determine one or more updated keywords from the one or more text strings, the one or more updated keywords being from the text strings having inventory word vectors with the highest cosine similarities to the input word vector.

9. The system of claim 1 , further including weighing further factors in addition to the inventory word vectors with the highest cosine similarities to the input word vector in order to analyze the calculated cosine similarities to determine the one or more keywords includes.

10. The system of claim 1 , further including accessing all item listings available over a predetermined time period in order to access a plurality of inventory item listings.

11. The system of claim 1 , further including reducing the text strings to words only in order to create the inventory word vectors.

12. The system of claim 11 , further including removing all special text characters and putting all words into lower case in order to reduce the text strings to words only.

13. The system of claim 1 , the one or more keywords having an associative context with the user input based on the cosine similarities rather than a context based on historical data analytics or direct manual listings by a secondary media content provider.

14. A method, comprising:

accessing, by a server computer, a plurality of inventory item listings accessible over a network, wherein each of the plurality of inventory item listings includes one or more text strings;

creating, by the server computer, inventory word vectors for at least some of the text strings of the plurality of inventory item listings;

receiving, by the server computer, a user input including an input word, wherein the user input is given by a user to access a primary media content;

creating, by the server computer, an input word vector for the input word;

calculating, by the server computer, cosine similarities between the input word vector and the inventory word vectors; and

analyzing, by the server computer, the calculated cosine similarities to determine one or more keywords from the one or more text strings, wherein the one or more keywords are from the text strings having inventory word vectors with the highest cosine similarities to the input word vector.

15. The method of claim 14 , further comprising:

associating the one or more keywords with a secondary media content; and

providing the secondary media content to be presented to the user.

16. The method of claim 15 , the primary media content being set of search results and the secondary media content is a targeted advertisement.

17. The method of claim 14 , the plurality of inventory item listings being listings from an online marketplace.

18. The method of claim 14 , further comprising:

creating revised inventory word vectors for at least some of the same text strings for which inventory word vectors were originally created;

calculating cosine similarities between the input word vector and the revised inventory word vectors; and

analyzing the calculated cosign similarities to determine one or more updated keywords from the one or more text strings, the one or more updated keywords being from the text strings having revised inventory word vectors with the highest cosine similarities to the input word vector.

19. The method of claim 18 , the revised inventory word vectors being created at least one day after the inventory word vectors were originally created.

20. A machine-readable storage device embodying non-tangible instructions that are executable by at least one processor to cause a machine to perform operations comprising:

accessing, by the at least one processor, a plurality of inventory item listings accessible over a network, the plurality of inventory item listings including one or more text strings;

creating by the at least one processor, inventory word vectors for at least some of the text strings of the plurality of inventory item listings;

receiving, by the at least one processor, a user input including an input word, the user input being input by a user to access a primary media content;

creating, by the at least one processor, an input word vector for the input word;

calculating, by the at least one processor, cosine similarities between the input word vector and the inventory word vectors; and

analyzing, by the at least one processor, the calculated cosine similarities to determine one or more keywords from the one or more text strings, the one or more keywords being from the text strings having inventory word vectors with the highest cosine similarities to the input word vector.