The technique offers a new approach to opinion-mining, which usually requires a team of dedicated surveyors and participants.
Understanding public opinion is a requirement for almost all organizations in the internet age, as campaigns, adverts, and products are launched to fit with current trends.
To meet this demand, a team of researchers at Hansung University in South Korea have developed an algorithm capable of summarizing online sentiments.
Published in the International Journal of Computational Vision and Robotics, the algorithm works by identifying and extracting representative posts, articles, and stories from a large swarth of content on a singular topic.
It is able to identify clearly the most important content, through activity and other metrics, and build a summarization from that central point. This could be a new approach to opinion-mining, which usually requires a team of dedicated surveyors and participants.
In the paper, Professor Jae-Young Chang provides proof of the algorithm’s capability, testing sentiment on popular movie reviews. By identifying the most important information online, the algorithm was able to come up with an accurate summary of online sentiment towards a movie.