A Topic Mapping-based framework to analyze textual risk reports from social media big data contents

Recruiting a decision support system that handles risks and monitors evidences will mitigate the destructive effects of risk. The main challenge of risk data is when it is unstructured, such as those metadata retrieved from posts published on social networks. The research gap focused in this study is analyzing unstructured risk data and extracting the embedded knowledge. We propose a Topic Map-based knowledge discovery system for analyzing unstructured risk data using the word embedding approach. We examined our proposed system on a dataset from some supply chain risk-related websites containing 325,965 sentences, 725,986 terms, and 52,396 unique terms. The experiment extracted 63 crisis knowledge propositions classified from this dataset using just three user-system interaction steps, which shows the model's high performance. The results reveal that the proposed model could be used effectively and efficiently in decision support systems to analyze unstructured data, especially for crises in the supply chain analysis.

Automated summary

This study focuses on analyzing unstructured risk data and extracting knowledge, and proposes a Topic Map-based knowledge discovery system using the word embedding approach. The experiment shows that the proposed system can effectively analyze unstructured data and performs well in supply chain analysis.