Early-stage pregnancy recognition on microblogs: Machine learning and lexicon-based approaches

Samer Muthana Sarsam; Ahmed Ibrahim Alzahrani; Hosam Al-Samarraie

doi:10.1016/j.heliyon.2023.e20132

Early-stage pregnancy recognition on microblogs: Machine learning and lexicon-based approaches

Heliyon. 2023 Sep 14;9(9):e20132. doi: 10.1016/j.heliyon.2023.e20132. eCollection 2023 Sep.

Authors

Samer Muthana Sarsam¹, Ahmed Ibrahim Alzahrani², Hosam Al-Samarraie^{3

4}

Affiliations

¹ School of Strategy and Leadership, Coventry University, Coventry, United Kingdom.
² Computer Science Department, Community College, King Saud University, Riyadh, 11437, Saudi Arabia.
³ School of Design, University of Leeds, Leeds, United Kingdom.
⁴ Centre for Instructional Technology and Multimedia, Universiti Sains Malaysia, Penang, Malaysia.

Abstract

Pregnancy carries high medical and psychosocial risks that could lead pregnant women to experience serious health consequences. Providing protective measures for pregnant women is one of the critical tasks during the pregnancy period. This study proposes an emotion-based mechanism to detect the early stage of pregnancy using real-time data from Twitter. Pregnancy-related emotions (e.g., anger, fear, sadness, joy, and surprise) and polarity (positive and negative) were extracted from users' tweets using NRC Affect Intensity Lexicon and SentiStrength techniques. Then, pregnancy-related terms were extracted and mapped with pregnancy-related sentiments using part-of-speech tagging and association rules mining techniques. The results showed that pregnancy tweets contained high positivity, as well as significant amounts of joy, sadness, and fear. The classification results demonstrated the possibility of using users' sentiments for early-stage pregnancy recognition on microblogs. The proposed mechanism offers valuable insights to healthcare decision-makers, allowing them to develop a comprehensive understanding of users' health status based on social media posts.

Keywords: Emotions; Pregnancy recognition; Sentiment analysis; Social media analysis.