As it seems “royal baby” fever will not disappear any soon, The Guardian (guardian.co.uk) allows users to filter any news about the royal family.  You can click the button says “Republican” on the front page and after some magic happens, all related stories are now filtered.
Although I only visit news websites when I have (really) free time, I usually don’t prefer it. There are lots of stories that really does not take my attention and finding the most relevant content takes some time. Well, you can be a fast internet user, but at least you should read a few words of the titles. When I start blogging in 2007, I was thinking how cool RSS is, but it leads the same problem again. You cannot know if you would like to read stories in your RSS list, even if you only follow stories (posts) of your favorite publisher (blogger).
Royal baby filter should give a hint to such website owners and newspapers. Although static filters may work for such specific events, the correct way to implement requires machine learning algorithms. Personalized news is actually an important topic that lots of people could interest. I am using Prismatic for a while, which allows you to select your interest and favorite publishers to deliver you a nice and readable updated list of social news. (An alternative for mobile users is Flipboard.) In near future, machine learning may be needed more to respond reader requests. Besides your common interests, you may like to read news based on your mood and daily preferences. Some developers are already on track to use machine learning to summarize news. 
There are some challenges for news aggregators like Flipboard and Prismatic to use machine learning, because they can only access restricted information about users such as which titles are clicked. Even this does not necessarily say that the user is interested with the subject. Maybe information such as leaving a comment, the total time spent in the page, jumping to another related news and bounce rate are useful for better predictions. However, these are not available to news aggregators. Therefore; having a limited data leads limited performance of machine learning tools.
And of course, you need to make a good categorization of news to make better suggestions. Interested readers can read Fabrizio Sebastiani’s paper, titled “Machine Learning in Automated Text Categorization” .
You can also find potential difficulties for utilizing machine learning to similar concepts in the paper written by Webb et al . I would like to hear any other sources that you can share with us. And lastly, what do you think about Royalist / Republican button of the Guardian?
 Indvik, Lauren. “‘The Guardian’ Lets Readers Switch Off Royal Baby News.” Mashable. N.p., 22 July 2013. Web. 23 July 2013.
 Lardinois, Frederic. “NewsRel Uses Machine Learning To Summarize News Stories And Put Them On A Map.” Tech Crunch. N.p., 28 Apr. 2013. Web. 23 July 2013.
 Sebastiani, Fabrizio. “Machine learning in automated text categorization.” ACM computing surveys (CSUR) 34.1 (2002): 1-47.
 Webb, Geoffrey I., Michael J. Pazzani, and Daniel Billsus. “Machine learning for user modeling.” User modeling and user-adapted interaction 11.1-2 (2001): 19-29.