Discovering long COVID symptom patterns: Association rule mining and sentiment analysis in social media tweets (Preprint)

Abstract

BACKGROUND The COVID-19 pandemic is a significant public health crisis that negatively affects human health and well-being. As a result of being infected with the Coronavirus, patients can experience long-term health effects, called long COVID. Multiple symptoms characterize this syndrome, and it is crucial to identify these symptoms as they may negatively impact patients’ day-to-day lives. Breathlessness, fatigue, and brain fog are the three main continuing and debilitating symptoms that long COVID patients have reported, often months after the onset of the COVID-19 disease. OBJECTIVE This study aimed to understand the patterns and behavior of long COVID symptoms, which is vital to improving our understanding of long COVID. METHODS Long COVID-19 related Twitter data were collected from 1 May 2020 to 31 December 2021. We used association rule mining techniques to identify frequent symptoms and describe symptom patterns among long COVID patients in Twitter social media discussions. RESULTS The most frequent symptoms in our study included brain fog, fatigue, breathing/lung issues, heart issues, flu symptoms, and depression. General pains, loss of smell and taste, cold, cough, chest pain, fever, headache, and arm pain emerged in two to six percent of long COVID patients. The highest confidence level-based detection successfully demonstrates the potential of association analysis and Apriori algorithm to establish patterns to detect 62 relationship rules among long COVID symptoms. CONCLUSIONS There are very active social media discussions that could support the growing understanding of the COVID-19 and its long-term impact. This enables a potential field of research to analyze the behavior of the long COVID syndrome. Exploratory data analysis was done to identify the symptoms and medical conditions related to long COVID discussions on the Twitter social media platform. Using Apriori algorithm-based association rules, we determined interesting relationships between symptoms.

A survey on the role of artificial intelligence in managing Long COVID

“…Regarding the approaches using models different from BERT, they are mostly employed for identifying Long COVID symptoms [see Matharaarachchi et al (2022) In terms of data, it is worth noting that most of the techniques are employed on textual data from Twitter. Among these techniques, three out of four which are based on association rule mining (Matharaarachchi et al, 2022), NLP and SVM (Banda et al, 2021), and Biterm Topic Modeling (Déguilhem et al, 2022), are adopted for Long COVID symptom identification and co-occurrence, while one of them (Miao et al, 2022) aims to describe the nature of Long COVID symptoms in terms of demographic features.…”

“…Differently from the previous approaches where classification is adopted, Matharaarachchi et al (2022) explored the trends and characteristics associated with Long COVID using the Apriori algorithm-based Association Rule Mining Technique. The focus of the study was to examine the common symptoms of patients with Long COVID and determine any correlations between them, using Twitter social media conversations as a reference.…”

“…Finally, Patterson et al (2021), Patel et al (2023), and Zhang et al (2023) adopted multiclass classification of laboratory data for Long COVID identification. By contrast, the approaches based on NLP in Table3are mostly related to Long COVID symptoms identification [seeBanda et al (2021),Déguilhem et al (2022),Matharaarachchi et al (2022),Miao et al (2022),Wang et al (2022), andZhu et al (2022)].…”