The Emotion in Text Analyzer: How to Visualize Its Output

P. 211–221.

Anastasia Kolmogorova, Alexander Kalinin

The article summarizes the results of the project conducted in the field of emotional text analysis. The project aim is to build up an analyzer able, according to the model of “Lövheim Cube”, to detect eight emotions in the Internet-texts in Russian. Having collected a labeled dataset and trained ML models, we faced the problem of the way to visualize the results of emotional text analysis. The main problem is due to the non-discrete approach in emotion in text assessment we are following in our research: we consider emotions as continuum and not as discrete categories. In this paper, we briefly describe the main logic of our work on the analyzer and focus on the different approaches to its output visual presentation. Finally, we demonstrate the design we have chosen to fit users’ expectations better and give an opening to the perspective of using the analyzer with such an interface.

Keywords: визуализация, data visualization, базовые эмоции, автоматическая обработка текстов, basic emotions, Natural Language Processing (NLP)

In book

Literature, Language and Computing: Russian Contribution

Springer, 2023.