NewsInn

NewsInn is an Artificial Intelligence Driven Algorithm that crawls news websites, parses and conglomerates news from the websites of major publications. The algorithm uses an opinion extraction procedure, a derivative of SentiWord, to generate a sentiment analysis for each news article.
More than 20 Major News Publications are used as source, scanned every half hour, and content is extracted from them. From this primary data, keywords are extracted, then every article gets a short summary automatically generated.
NewsInn uses an opinion extraction algorithm that quantifies the positive or negative outlook of every news story. This algorithm was proven to predict stock market indices with an accuracy of up to 66.7%.<ref name="Academia"/> The study was published in MacroThink's Business and Economic Research Journal<ref name="MacroThink"/>
Description of Algorithm
The initial process parses through the web version of a number of news publications. The HTML tags are stripped and a temporary, text version is stored. The Opinion extraction algorithm is the first to take effect, comparing each word in the article to a known dictionary and attributing a "Positive" or "Negative" score.<ref name="Academia"/><ref name="MacroThink"/> A simple summarisation algorithm then parses the text and creates a summary.
Keywords are extracted based on frequency in the text, if they are the names of people, geographical locations or entity names. A Machine Learning Algorithm is in place, lowering or raising the importance of certain keywords. On the last step, articles that have similar keywords are grouped together to form a single news story.<ref name="AboutUs"/>
A Social Media algorithm automatically posts the latest news on Twitter and Facebook, after generating hashtags based on keywords.
 
< Prev   Next >