Analyzing Moroccan Tweets to Extract Sentiments Related to the Coronavirus Pandemic: A New Classification Approach

摘要

At the end of 2019, the world has known the covid-19 crisis that negatively affected the health, economic, social, and psychological status of people. Since the beginning of this crisis, users express their ideas, opinions, and sentiments about the coronavirus on all social networks such as Facebook, Twitter, Instagram, etc. For example, until May 8th, 2020, the number of tweets published on Twitter is equal to 628,809,016. In this paper, our proposed method analyzes and classifies covid-19 tweets published in morocco for extracting sentiments. Our approach uses the advantages of new proposed tweets features using a dictionary-based approach and a Python library for developing a new recommendation approach. As Experiments, Our proposed approach outperforms the well-known machine learning classifiers. We find also that based on the epidemiological situation in morocco, the sentiments of Moroccan users changed.