RT info:eu-repo/semantics/article
T1 Machine learning and natural language processing to assess the emotional impact of influencers’ mental health content on Instagram
A1 Merayo Álvarez, Noemí
A1 Ayuso Lanchares, Alba
A1 González Sanguino, Teresa Clara
K1 Mental health
K1 Sentiment analysis
K1 Emotions
K1 Machine learning
K1 Social networks
K1 Instagram
AB Background: This study aims to examine, through artificial intelligence, specificallymachine learning, the emotional impact generated by disclosures about mentalhealth on social media. In contrast to previous research, which primarily focused onidentifying psychopathologies, our study investigates the emotional response tomental health-related content on Instagram, particularly content created byinfluencers/celebrities. This platform, especially favored by the youth, is the stagewhere these influencers exert significant social impact, and where their analysis holdsstrong relevance. Analyzing mental health with machine learning techniques onInstagram is unprecedented, as all existing research has primarily focused on Twitter.Methods: This research involves creating a new corpus labelled with responses tomental health posts made by influencers/celebrities on Instagram, categorized byemotions such as love/admiration, anger/contempt/mockery, gratitude,identification/empathy, and sadness. The study is complemented by modelling a setof machine learning algorithms to efficiently detect the emotions arising when facedwith these mental health disclosures on Instagram, using the previous corpus.Results: Results have shown that machine learning algorithms can effectively detectsuch emotional responses. Traditional techniques, such as Random Forest, showeddecent performance with low computational loads (around 50%), while deep learningand Bidirectional Encoder Representation from Transformers (BERT) algorithmsachieved very good results. In particular, the BERT models reached accuracy levelsbetween 86–90%, and the deep learning model achieved 72% accuracy. These resultsare satisfactory, considering that predicting emotions, especially in social networks, ischallenging due to factors such as the subjectivity of emotion interpretation, thevariability of emotions between individuals, and the interpretation of emotions indifferent cultures and communities.Discussion: This cross-cutting research between mental health and artificialintelligence allows us to understand the emotional impact generated by mental healthcontent on social networks, especially content generated by influential celebritiesamong young people. The application of machine learning allows us to understandthe emotional reactions of society to messages related to mental health, which is highly innovative and socially relevant given the importance of the phenomenon insocieties. In fact, the proposed algorithms’ high accuracy (86–90%) in social contextslike mental health, where detecting negative emotions is crucial, presents a promisingresearch avenue. Achieving such levels of accuracy is highly valuable due to thesignificant implications of false positives or false negatives in this social context.
SN 2376-5992
YR 2024
FD 2024
LK https://uvadoc.uva.es/handle/10324/75288
UL https://uvadoc.uva.es/handle/10324/75288
LA eng
NO PeerJ Computer Science, 2024, vol. 10, e2251
DS UVaDOC
RD 26-abr-2025