Abstract:Emotion mining has become a crucial tool for understanding human emotions during disasters, leveraging the extensive data generated on social media platforms. This paper aims to summarize existing research on emotion mining within disaster contexts, highlighting both significant discoveries and persistent issues. On the one hand, emotion mining techniques have achieved acceptable accuracy enabling applications such as rapid damage assessment and mental health surveillance. On the other hand, with many studies adopting data-driven approaches, several methodological issues remain. These include arbitrary emotion classification, ignoring biases inherent in data collection from social media, such as the overrepresentation of individuals from higher socioeconomic status on Twitter, and the lack of application of theoretical frameworks like cross-cultural comparisons. These problems can be summarized as a notable lack of theory-driven research and ignoring insights from social and behavioral sciences. This paper underscores the need for interdisciplinary collaboration between computer scientists and social scientists to develop more robust and theoretically grounded approaches in emotion mining. By addressing these gaps, we aim to enhance the effectiveness and reliability of emotion mining methodologies, ultimately contributing to improved disaster preparedness, response, and recovery.
Keywords: emotion mining, sentiment analysis, natural disasters, psychology, technological disasters
What problem does this paper attempt to address?
The main problems that this paper attempts to solve are the current methodological problems of sentiment mining in disaster situations, specifically including:
1. **Lack of theory - driven research**: Although sentiment mining techniques have made significant progress in disaster situations, many studies rely too much on data - driven methods and ignore the theoretical frameworks of social and behavioral sciences. This has led to arbitrariness in sentiment classification and the neglect of factors such as cultural differences.
2. **Arbitrariness in sentiment classification**: Many studies do not base on existing sentiment theories (such as discrete emotion theory, dimensional emotion theory, and appraisal theory) when classifying emotions, but arbitrarily select emotion categories, which makes it difficult to compare and synthesize the results between different studies.
3. **Neglect of cross - cultural differences**: Although sentiment mining research covers multiple cultures and languages, there is a lack of systematic cross - cultural comparison, and it is impossible to draw conclusions about the universality and particularity of emotional responses in different cultural backgrounds.
4. **Impact of data bias and false information**: There are obvious biases in social media data, such as the over - representation of people with high socioeconomic status on Twitter and the widespread spread of false information, which may distort the real emotional state of the public.
In order to solve these problems, the paper calls for interdisciplinary cooperation between computer scientists and social scientists to develop more robust and theoretically - based sentiment mining methods, thereby improving the effectiveness and reliability of disaster preparedness, response, and recovery.
### Formula example
When discussing sentiment classification, the paper mentions the discrete emotion theory, in which each basic emotion has its own unique characteristics. For example, the six basic emotions proposed by Ekman (happiness, fear, anger, sadness, surprise, and disgust) can be expressed as:
\[
E=\{happiness, fear, anger, sadness, surprise, disgust\}
\]
And in some studies, the sentiment classification may be:
\[
E' = \{positive, neutral, angry, anxious, fearful, sad\}
\]
This classification method lacks a theoretical basis, resulting in difficulty in comparison and synthesis between different studies.