Automatic article enrichment by social media trends

A trend-content-presentation server comprising:

• a client interface for communicating, via a network, with at least one client terminal hosting a browser displaying a web document;
• a trend datastore that hosts a trend dataset storing one or more trends each of which appears in one or more web documents;
• a program store storing code;
• and
• a processor coupled to the client interface, the trend datastore, and the program store the stored code comprising:
• code to receive, from a browser on a client terminal, a designation of the web document for analysis, the designation comprising a Universe Resource Locator (URL) where the web document is located on a web server after the browser obtains the web document from the web server and displays the web document on the client terminal, the web document comprising a script configured to cause the browser to generate the designation and send to the trend-content-presentation server the designation that comprises the URL, the web server being configured to add the script to the web document;
• code to respond to the designation by using the received URL to analyze the web document and identify a trend related to content of the web document, the trend being represented by a representative term;
• code to send, to a plurality of social network servers, a plurality of first requests, each first request comprising the representative term;
• code to receive, from the plurality of social network servers, a plurality of first user uploaded media content items in response to the plurality of first requests;
• code to extract terms from the plurality of first user uploaded media content items;
• code to identify a set of identified terms correlated with the trend by applying a statistical classifier to the extracted terms, the statistical classifier being configured to:
• assign similarity values to one or more terms in the set of identified terms based on similarity to the trend, the one or more terms comprising a first term and a second term, the first term clustering closer to the trend than the second term, and the first term is assigned a higher similarity value than the second term;
• and
• identify the set of identified terms from the extracted terms based on the similarity values and a similarity requirement, the similarity requirement being a threshold value that is dynamically adjusted upward in response to the set of identified terms being smaller than a predefined size or downward in response the set of identified terms being larger than the predefined size;
• code to send, to the plurality of social network servers, a plurality of second requests, each second request comprising at least one identified term from the set of identified terms;
• code to receive, from the plurality of social network servers, a plurality of second user uploaded media content items in response to the plurality of second requests;
• code to select at least one media content item of the plurality of second user uploaded media content items based on at least one selection parameter, the at least one selection parameter including language similarity between the at least one identified term and the at least one media content item, the selecting of the at least one media content item of the plurality of second user uploaded media content items including:
• ranking the plurality of second user uploaded media content items according to at least one characteristic of the plurality of second user uploaded media content items that is independent of the trend, the at least one characteristic being a member selected from a group consisting of:
• image quality;
• and
• storage size;
• and
• determining, from the ranked plurality of second user uploaded media content items, the at least one media content item that has at least one of image quality suitable for presentation on the client terminal or storage size suitable for presentation on the client terminal;
• code to send, to the browser on the client terminal, the at least one media content item to a presentation on the client terminal, the script of the web document being further configured to cause the presentation of the at least one media content item to be simultaneously presented with the web document being presented by the browser, the presentation being implemented as an overlay that is displayed over and covering a part of the web document;
• and
• code to receive, from the browser on the client terminal, a request from a client terminal user to add feedback to the at least one media content item, the feedback being provided by the client terminal user through the presentation, the feedback comprising a comment regarding the at least one media content item, the feedback being associated with an uploading user that uploaded the at least one media content item to one of the plurality of social network servers.