This article presents an original method for predicting the development of incidents in real time at the initial stage of a fire. It consists of three levels, each performing calculations in a specific direction. The first level assesses the degree of trust in information published by users of the analyzed social network. The second level assesses the degree of trust in information published by communities of the analyzed social network. Calculations at the first and second levels are performed using an algorithm for identifying reliable sources of information about fire-related incidents. The third level analyzes text content generated by users and communities of the analyzed social network. This is accomplished using a text analysis algorithm. The developed method synthesizes computational mechanisms within the algorithm and predicts fire development dynamics. The use of the proposed scientific tool in the work of units of EMERCOM of Russia will hypothetically improve the efficiency of emergency services, thereby saving lives and minimizing property damage.
Deineka et al. (Fri,) studied this question.