JMIR Publications

ABSTRACT

Background:

Public engagement is a key element for mitigating pandemics, and a good understanding of public opinion could help to encourage the successful adoption of public health measures by the population. Deep learning has been more and more applied to the analysis of text from social networks. However, most of developed approaches can only capture concerns or sentiments alone, but not both together.

Objective:

Here, we aimed to develop a new approach, based on deep neural networks, for simultaneously capturing public concerns and sentiments, and applied it to tweets sent just after the announcement of the SARS-CoV-2 pandemic by the WHO.

Methods:

Tweets were collected, preprocessed, and split 80/20 into training/validation sets. We combined lexicons and convolutional neural networks for improving sentiment prediction. The trained model was able to capture simultaneously the weighted words associated to a predicted sentiment intensity score. These outputs were then visualized via an interactive and customizable web interface based on a word-cloud representation. Using word cloud analysis, we captured the main concerns for extreme positive and negative sentiment intensity scores.

Results:

In reaction to the announcement of the pandemic by the WHO, six negative and five positive issues were discussed on Twitter. Twitter users seemed to be worried about the international situation, the economic consequences, and the medical situation. Conversely, they seemed to be satisfied with the commitment of medical and social workers and with the collaboration between people.

Conclusions:

We propose a new method based on deep neural networks for simultaneously extracting public concerns and sentiments from tweets. This method could be helpful for monitoring public opinion during crises such as pandemics.