Multiple criteria and statistical sentiment analysis on flooding

Abstract

Abstract The World Bank lists flooding as one of the main pressures on a community. Flooding can affect development prospects and potentially reverse decades of progress in the alleviation of poverty and in development. Flooding-related information usually involves many stakeholders, objects, and significant details. We examined linkage between the density of 506 flood management keywords from documents found via Google Search and 32 macroenvironment indicators for 76 countries, developing 506 neural networks models. These models show that flood management keywords are interconnected with the environmental, social, economic, political, and cultural dimensions of the examined countries. The models demonstrate that improvements in a country’s sustainability and performance metrics are followed by an increase in flood keyword use. Microsoft Azure AI and ChatGPT were used to generate abstractive summaries for the 100 most dense and statistically significant flood management keywords and 28 macroenvironment indicators for the individual countries included in our analysis, and for the world. We reduced the number of flood keywords density variables from 506 to 100 by selecting only the most relevant ones in order to generate abstractive summaries. Our Web of Science Sentiment Analysis Articles Model and Map of the World demonstrates that nations with an unfavorable macroenvironment published disproportionately fewer papers on sentiment analysis. This research benefits stakeholders by providing comprehensive and holistic information and data analysis, focusing on evidence-based flooding information.

Keywords

• Sentiment analysis
• Floods
• Holistic approach
• World models
• World map
• Multiple criteria and statistical analysis