سحر يكه فلاح

A crisis is a situation in which a drastic change occurs with a certain probability and leads to undesirable and unpredictable consequences. A sustained crisis is a type of crisis that often lasts for weeks, months, or years, including air pollution, drought, AIDS, the coronavirus pandemic, climate change, food shortages, and explosions. nuclear These types of crises have severe human, financial, and environmental consequences for society. In such a situation, due to the time limit, uncertainty, and the effects of the crisis, it becomes difficult to make a decision, and any wrong decision leads to irreparable damages. The research conducted in the field of sustainable crisis management includes research that has developed decision support systems and recommender systems for sustainable crisis management and researches in which the Twitter social network has been used to help decision makers in sustained crisis conditions. In this research, for better management of sustained crises, a general architecture based on a data-driven decision support system was introduced. The data-driven decision support system collects data related to the sustained crises over time from the Twitter social network. To eva‎luate the proposed architecture, a case study was conducted on people with underlying medical conditions and pregnancy to facilitate the decision to inject the COVID-19 vaccine. The dataset used includes the wisdom of the crowd regarding the COVID-19 vaccine injection, which was collected through the Twitter social network. The collected data set was labeled using a labeled data set and the BERT algorithm with the labels of favorable, unfavorable, and neutral positions regarding the injection of the COVID-19 vaccine. C4.5 decision tree algorithm with a 10-fold cross-validation method was used for decision modeling. Algorithm C4.5 achieved prediction precision of 0.88 and 0.55 respectively for opposing and agreeing positions towards COVID-19 vaccine injection. Then, with the help of an expert's knowledge, the COVID-19 vaccine recommendations were created and provided to users through the user interface. The results of the research show that social network data plays a significant role in reducing human decision-making errors in the management of sustained crises.

سامانه هاي تصميم يار داده محور , سامانه توصيه گر , تحليل پيامدها , شبكه هاي اجتماعي , بحران هاي پايدار

Data-driven decision support systems , Recommender system , Consequence analysis , Social networks , Sustained crises

Sahar Yekefallah

Mohammad Reza Kangavari