On the algorithmic role of synonyms and keywords in analytics for catastrophic events

We study statistical issues related to the use of words in messages on SNs, as related to specific potentially catastrophic events, such as fires and earthquakes. The analysis investigates the frequency of the words in the messages of interest, moreover their temporal correlation, and aims to determine a reasonable number of keywords for the searches of the events, especially in relation with analytics and searches on SNs for the reflectance of those events. A special attention is given to synonyms.