The Impact of News Values and Linguistic Style on the Popularity of Headlines On Twitter and Facebook

AAAI Conferences

A large proportion of audiences read news online, often accessing news articles through social media like Facebook or Twitter. A distinguishing characteristic of news on social media is that the most prominent (and often the only visible) part of the news article is the headline. We investigate the impact of headline characteristics, including journalistic concepts of news values and linguistic style, on the article's social media popularity. Using a large corpus of headlines from The Guardian and New York Times we derive these features automatically and correlate with social media popularity on Twitter and Facebook. We found most of them to have a significant effect and that their impact differs between the two social media and between news outlets. Further investigation with a crowdsourced study confirms that news values and style influence the audiences' decisions to click on a headline.


The Divided Kingdom: a machine learning analysis on the Brexit result MonkeyLearn Blog

#artificialintelligence

Today was a day for the history books. The UK has voted to leave the European Union and opened a deep crack in the heart of Europe. As a consequence of this result, Prime Minister David Cameron will step down by October urging for a fresh leadership. At this point nobody knows the repercussions of these results. Will the Brexit hurt the economy of the UK and ignite a new recession?


The Divided Kingdom: a machine learning analysis on the Brexit result MonkeyLearn Blog

#artificialintelligence

Today was a day for the history books. The UK has voted to leave the European Union and opened a deep crack in the heart of Europe. As a consequence of this result, Prime Minister David Cameron will step down by October urging for a fresh leadership. At this point nobody knows the repercussions of these results. Will the Brexit hurt the economy of the UK and ignite a new recession?


From Monologue to Dialogue: Natural Language Generation in OVIS

AAAI Conferences

This paper describes how a language generation system that was originally designed for monologue generation, has been adapted for use in the OVIS spoken dialogue system. To meet the requirement that in a dialogue, the system's utterances should make up a single, coherent dialogue turn, several modifications had to be made to the system.


Social Media Sentiment Analysis, and Soccer Meltwater

#artificialintelligence

Before delving into the nitty gritty of exactly how sentiment analysis works, let's break the concept down into something a little more tangible, shall we. Have you ever wondered what the South African public thought about, let's say, Iceland's football team defeating England in the Euro 2016? Well, that right there my friends, is why sentiment analysis software exists – to make vast quantities of data easily understandable at a glance. Think of it like a snapshot of the emotional response to a given topic. You might be asking yourself, but what about online surveys and polls, isn't that their purpose?