Personalized search filter and notification system
One or more techniques and/or systems for sending push notifications of content items to client devices are provided herein. For example, an input received from a user can be expanded to obtain an expanded user interest. Content items from a content source can be filtered based upon the expanded user interest to obtain a set of filtered content items. A push notification can be constructed to comprise one or more of the filtered content items from the set of filtered content items. The push notification can be sent to a client device of the user for display as a device alert notification. In an example, the filtered content items, within the push notification, may be ranked based upon a ranking metric.
1. A method, comprising:
receiving an input associated with a user;
expanding the input to obtain an expanded user interest, the expanding comprising:
identifying a term within the input;
identifying a content item based upon the input;
searching the content item for one or more co-occurring terms that co-occur with the term within the content item;
identifying a co-occurring term as one of the one or more co-occurring terms that co-occur with the term within the content item;
obtaining a plurality of content items based upon the input;
processing the plurality of content items obtained based upon the input to identify at least one of one or more categories or one or more topics, wherein the processing comprises determining that a majority of the plurality of content items obtained based upon the input are associated with at least one of a category or a topic; and
including the co-occurring term and at least one of the category or the topic in the expanded user interest; and
filtering content items from a content source based upon the expanded user interest to obtain a set of filtered content items.
2. The method of claim 1 , comprising ranking the set of filtered content items based upon a ranking metric to create a ranked set of filtered content items.
3. The method of claim 1 , comprising providing for display a notification based upon the set of filtered content items.
4. The method of claim 2 , comprising:
maintaining filtered content items, of the ranked set of filtered content items, within a cache;
defining a cache expiration period for the cache; and
responsive to expiration of the cache expiration period for a filtered content item, removing the filtered content item from the cache.
5. The method of claim 4 , the cache expiration period comprising between about a 6 hour expiration to about a 48 hour expiration.
6. The method of claim 2 , a filtered content item, within the ranked set of filtered content items, comprising at least one of:
a news article, a sports statistic, product pricing information, a travel deal, a social network post, or a stock update.
7. The method of claim 1 , the input comprising:
a search query comprising at least one of a second term or a phrase.
8. The method of claim 1 , the co-occurring term comprising a noun.
9. The method of claim 2 , the ranking comprising:
using at least one of a Gradient Boosting Decision Tree (GBDT) or an XQuery Data Model (XDM) to rank the set of filtered content items.
10. The method of claim 1 , comprising:
assigning a rating to a filtered content item based upon user feedback.
11. A system, comprising:
a processor; and
memory comprising processor-executable instructions that when executed by the processor cause the processor to implement:
a receiving component configured to:
receive an input associated with a user;
an input expansion component configured to:
expand the input to obtain an expanded user interest, the expanding comprising:
performing a search query using the input;
obtaining a plurality of content items based upon the search query; and
processing the plurality of content items obtained based upon the search query to identify at least one of one or more categories or one or more topics, wherein the processing comprises at least one of:
processing two or more content items, of the plurality of content items, identified based upon the input through a categorizer to (i) determine a category and (ii) include the category in the expanded user interest responsive to determining that a majority of the two or more content items processed through the categorizer are associated with the category; or
processing two or more content items, of the plurality of content items, identified based upon the input through a topic model to (i) determine a topic and (ii) include the topic in the expanded user interest responsive to determining that a majority of the two or more content items processed through the topic model are associated with the topic; and
a filtering component configured to:
filter content items from a content source based upon the expanded user interest to obtain a set of filtered content items.
12. The system of claim 11 , the filtering component configured to:
maintain filtered content items within a cache;
define a cache expiration period for the cache; and
responsive to expiration of the cache expiration period for a filtered content item, remove the filtered content item from the cache.
13. The system of claim 11 , wherein the processor-executable instructions cause the processor to implement:
a feedback component configured to:
assign a rating to a filtered content item based upon user feedback.
14. The system of claim 11 , the input expansion component configured to:
identify a term within the input.
15. A non-transitory machine readable medium having stored thereon processor-executable instructions that when executed cause performance of operations, the operations comprising:
receiving an input associated with a user;
expanding the input to obtain an expanded user interest, the expanding comprising:
performing a search query using the input;
obtaining a plurality of content items based upon the search query;
processing the plurality of content items obtained based upon the search query to identify at least one of one or more categories or one or more topics, wherein the processing comprises determining that a majority of the plurality of content items obtained based upon the search query performed using the input are associated with at least one of a category or a topic; and
including at least one of the category or the topic in the expanded user interest; and
filtering content items from a content source based upon the expanded user interest to obtain a set of filtered content items.
16. The non-transitory machine readable medium of claim 15 , the operations comprising ranking the set of filtered content items based upon a ranking metric to create a ranked set of filtered content items.
17. The non-transitory machine readable medium of claim 15 , the operations comprising sending a notification to an operating system based upon the set of filtered content items.
18. The non-transitory machine readable medium of claim 15 , the expanding comprising:
determining that the majority of the plurality of content items are associated with the category; and
including the category in the expanded user interest.
19. The non-transitory machine readable medium of claim 15 , the expanding comprising:
determining that the majority of the plurality of content items are associated with the topic; and
including the topic in the expanded user interest.
20. The non-transitory machine readable medium of claim 15 , the operations comprising:
assigning a rating to a filtered content item based upon user feedback.