Hamed Vahdat-Nejad - Perlab, Faculty of Electrical and Computer Engineering University of Birjand Birjand, Iran
Faezeh Azizi - Perlab, Faculty of Electrical and Computer Engineering University of Birjand Birjand, Iran
Mahdi Hajiabadi - Perlab, Faculty of Electrical and Computer Engineering University of Birjand Birjand, Iran
Fatemeh Salmani - Perlab, Faculty of Electrical and Computer Engineering University of Birjand Birjand, Iran
Sajedeh Abbasi - Perlab, Faculty of Electrical and Computer Engineering University of Birjand Birjand, Iran
Mohadese Jamalian - Perlab, Faculty of Electrical and Computer Engineering University of Birjand Birjand, Iran
Reyhane Mosafer - Perlab, Faculty of Electrical and Computer Engineering University of Birjand Birjand, Iran
Hamideh Hajiabadi - Perlab, Faculty of Electrical and Computer Engineering University of Birjand Birjand, Iran
Wathiq Mansoor - Department of Electrical Engineering University of Dubai Dubai, UAE

The outbreak of the COVID-۱۹ in ۲۰۲۰ and lack of an effective cure caused psychological problems among humans. This has been reflected widely on social media. Analyzing a large number of English tweets posted in the early stages of the pandemic, this paper addresses three psychological parameters: fear, hope, and depression. The main issue is the extraction of the related tweets with each of these parameters. To this end, three lexicons are proposed for these psychological parameters to extract the tweets through content analysis. A lexicon-based method is then used with GEO Names (i.e. a geographical database) to label tweets with country tags. Fear, hope, and depression trends are then extracted for the entire world and ۳۰ countries. According to the analysis of results, there is a high correlation between the frequency of tweets and the official daily statistics of active cases in many countries. Moreover, fear tweets dominate hope tweets in most countries, something which shows the worldwide fear in the early months of the pandemic. Ultimately, the diagrams of many countries demonstrate unusual spikes caused by the dissemination of specific news and announcements.

natural language processing, emotion analysis, knowledge extraction, data mining.