Argus: Sentiment analysis developed with UZH

Sentiment analysis can be used to automatically determine the mood of print, online and social media messages within a very short time. The function is integrated into the all-in-one platform Argusavenue and is available to all customers with immediate effect.

Together with ICL, Argus has spent a year developing a solution for the automatic evaluation of text content in German, French and English. This expands Argus' media intelligence offering and enables customers to gain a quick overview of the sentiment of a large number of articles. A distinction is made between "positive", "neutral", "negative" and "controversial". In addition to the evaluation of the entire article, the analysis can also be narrowed down to individual keywords (detailed evaluation). This automatic evaluation is integrated into the new all-in-one platform Argusavenue. In combination with the newsletter module, automatic alerts can be set up for negative reports in order to be able to react promptly to reputational risks.

While a great deal has already been achieved in the field of sentiment in English, there is still a lot of catching up to do, especially in German and French. Closing this gap is an important goal of the joint project. "As a Swiss company, it is important to us that we initially focus on the languages German, French and English when it comes to the machine analysis of texts. In the University of Zurich, we have found the ideal partner to help us make progress in this area," explains Linard Campell, Head of Product/Business Development at Argus.

Manfred Klenner, ICL project manager: "The focus of the development was the creation of a new multilingual sentiment analysis. The vision was a system that can not only differentiate between simple positive and negative evaluations as usual. New filter functions should also make it possible to distinguish between positive and negative facts on the one hand and moral and emotional 'tones' on the other."