The ability to measure audiences during television commercial breaks versus the overall program has created a new challenge for broadcasters. Advertising revenues provide the foundation of a networks ability to invest in programming, including their news product. The ability to maintain audiences during commercial breaks increases the revenue potential and helps maintain a network’s ability to produce independent journalism. With high commercial clutter currently, viewership declines during commercial breaks is high, reducing the amount of revenue network’s earn. Accurate estimation of the extent to which viewership drops when a commercial is presented can have many benefits:
Artificial Neural Network algorithms (ANNs) have been applied to many predictive and forecasting problems and often outperform traditional regression tasks. Recent developments in Deep Learning algorithms have found application in consumer devices, addressing difficult pattern recognition problems that arise in image processing and speech recognition. The proposed work will apply such algorithms to the prediction of reduction in commercial viewership.
Professors Beth Egan and Fiona Chew of the Newhouse School, and Chilukuri Mohan of the College of Engineering and Computer Science, combine expertise from advertising, media studies and machine learning. Much of the research data provided by Comscore.
This study has received a 2019 CUSE Grant.