New Software Promises to Predict TV Program Audience

A consortium including ad management software firm SintecMedia has developed an advanced machine learning algorithm which it says will forecast future viewing of television programs.

The consortium, founded by the Israeli Science Ministry, includes government, academic and business experts in data science and technology. As part of the five-year, $25m initiative, the consortium devised a system using neural networks and natural language processing to predict GRP ratings. This process uses thousands of disparate types of information to create a forecast - including historical performance data, performance of similar programming, social media data from sites such as Twitter and Facebook, trending topics on search, and even weather and holiday data. According to SintecMedia, this formula not only predicts the success and size of viewership of upcoming, advertised shows, it can also be used to predict the future success of similar programming.

Roy Gelbard (pictured), Professor of Information and Knowledge Systems at Bar-Ilan University, comments: 'We are very encouraged by the success of our initial tests. This project has the very real potential to save media companies millions in lost advertising revenue, while significantly improving their current advertising operational efficiency. This breakthrough will not only help media companies sell more accurately on TV, but takes them much closer to selling at scale across TV and digital media'.

Web site: www.sintecmedia.com .