Automated Analysis of Emotions in News Texts

The paper discusses some pilot results of a research aimed at creating a model for emotion analysis in news texts that can identify and categorize emotions into a wide set of predefined categories. The developed model combines convolutional neural networks and long short-term memory networks, supported with an attention mechanism, to achieve high accuracy in emotion analysis. The model has been trained and validated, showing high accuracy and efficiency in emotion classification. The experimental results demonstrate that our model can reliably predict the emotional content of news texts and can also be used to predict the impact of news, presented in an appropriately emotionally intensive manner, on public opinion.