Dataset Information

Depressive Emotion Detection and Behavior Analysis of Men Who Have Sex With Men via Social Media.

ABSTRACT: Background:A large amount of evidence has indicated an association between depression and HIV risk among men who have sex with men (MSM), but traditional questionnaire-based methods are limited in timely monitoring depressive emotions with large sample sizes. With the development of social media and machine learning techniques, MSM depression can be well monitored in an online and easy-to-use manner. Thereby, we adopt a machine learning algorithm for MSM depressive emotion detection and behavior analysis with online social networking data. Methods:A large-scale MSM data set including 664,335 users and over 12 million posts was collected from the most popular MSM-oriented geosocial networking mobile application named Blued. Also, a non-MSM Benchmark data set from Twitter was used. After data preprocessing and feature extraction of these two data sets, a machine learning algorithm named XGBoost was adopted for detecting depressive emotions. Results:The algorithm shows good performance in the Blued and Twitter data sets. And three extracted features significantly affecting the depressive emotion detection were found, including depressive words, LDA topic words, and post-time distribution. On the one hand, the MSM with depressive emotions published posts with more depressive words, negative words and positive words than the MSM without depressive emotions. On the other hand, in comparison with the non-MSM with depressive emotions, the MSM with depressive emotions showed more significant depressive symptoms, such as insomnia, depressive mood, and suicidal thoughts. Conclusions:The online MSM depressive emotion detection using machine learning can provide a proper and easy-to-use way in real-world applications, which help identify high-risk individuals at the early stage of depression for further diagnosis.

SUBMITTER: Li Y

PROVIDER: S-EPMC7456911 | biostudies-literature | 2020

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Depressive Emotion Detection and Behavior Analysis of Men Who Have Sex With Men <i>via</i> Social Media.

Li Yong Y Cai Mengsi M Qin Shuo S Lu Xin X

Frontiers in psychiatry 20200814

<h4>Background</h4>A large amount of evidence has indicated an association between depression and HIV risk among men who have sex with men (MSM), but traditional questionnaire-based methods are limited in timely monitoring depressive emotions with large sample sizes. With the development of social media and machine learning techniques, MSM depression can be well monitored in an online and easy-to-use manner. Thereby, we adopt a machine learning algorithm for MSM depressive emotion detection and ...[more]

PMID: 32922323

Similar Datasets

Project description:BACKGROUND:Gay, bisexual, and other cisgender men who have sex with men (GBMSM) are disproportionately affected by the HIV pandemic. Traditionally, GBMSM have been deemed less relevant in HIV epidemics in low- and middle-income settings where HIV epidemics are more generalized. This is due (in part) to how important population size estimates regarding the number of individuals who identify as GBMSM are to informing the development and monitoring of HIV prevention, treatment, and care programs and coverage. However, pervasive stigma and criminalization of same-sex practices and relationships provide a challenging environment for population enumeration, and these factors have been associated with implausibly low or absent size estimates of GBMSM, thereby limiting knowledge about the dynamics of HIV transmission and the implementation of programs addressing GBMSM. OBJECTIVE:This study leverages estimates of the number of members of a social app geared towards gay men (Hornet) and members of Facebook using self-reported relationship interests in men, men and women, and those with at least one reported same-sex interest. Results were categorized by country of residence to validate official size estimates of GBMSM in 13 countries across five continents. METHODS:Data were collected through the Hornet Gay Social Network and by using an a priori determined framework to estimate the numbers of Facebook members with interests associated with GBMSM in South Africa, Ghana, Nigeria, Senegal, Côte d'Ivoire, Mauritania, The Gambia, Lebanon, Thailand, Malaysia, Brazil, Ukraine, and the United States. These estimates were compared with the most recent Joint United Nations Programme on HIV/AIDS (UNAIDS) and national estimates across 143 countries. RESULTS:The estimates that leveraged social media apps for the number of GBMSM across countries are consistently far higher than official UNAIDS estimates. Using Facebook, it is also feasible to assess the numbers of GBMSM aged 13-17 years, which demonstrate similar proportions to those of older men. There is greater consistency in Facebook estimates of GBMSM compared to UNAIDS-reported estimates across countries. CONCLUSIONS:The ability to use social media for epidemiologic and HIV prevention, treatment, and care needs continues to improve. Here, a method leveraging different categories of same-sex interests on Facebook, combined with a specific gay-oriented app (Hornet), demonstrated significantly higher estimates than those officially reported. While there are biases in this approach, these data reinforce the need for multiple methods to be used to count the number of GBMSM (especially in more stigmatizing settings) to better inform mathematical models and the scale of HIV program coverage. Moreover, these estimates can inform programs for those aged 13-17 years; a group for which HIV incidence is the highest and HIV prevention program coverage, including the availability of pre-exposure prophylaxis (PrEP), is lowest. Taken together, these results highlight the potential for social media to provide comparable estimates of the number of GBMSM across a large range of countries, including some with no reported estimates.

Project description:BACKGROUND:Many interventions find that social media engagement with health promotion materials can translate into behavioral changes. However, only a few studies have examined the ways in which specific actions on various social media platforms are correlated with health behaviors. OBJECTIVE:The objective of this study was to examine the association between social media use and HIV testing behaviors among Chinese men who have sex with men (MSM). METHODS:In July 2016, a Web-based survey was conducted to recruit MSM in 8 Chinese cities through Blued (Blue City Holdings Ltd.), the world's largest gay mobile phone app. Data on sociodemographic variables, social media use platforms and behaviors, sexual behaviors, and HIV testing histories were collected. HIV testing-related social media use was defined as having ever engaged with HIV testing content on social media, which was further divided into observing (ie, receiving), endorsing (eg, liking and sharing), and contributing (eg, posting or commenting on HIV testing materials). Confirmatory factor analysis (CFA) was conducted to determine the best division of HIV testing-related social media use. Univariate and multivariable logistic regressions were used to examine the association between HIV testing-related social media use and HIV testing behaviors. RESULTS:A total of 2105 individuals participated in the survey. Among them, 46.75% (984) were under the age of 24 years, 35.43% (746) had high school education or less, and 47.74% (587) had condomless sex in the last 3 months. More than half of the respondents (58.14%, 1224/2105) reported HIV testing-related social media use. Additionally, HIV testing-related social media use, especially on multifunctional platforms such as WeChat, was found to be associated with recent HIV testing (adjusted odds ratio [aOR] 2.32, 95% CI 1.66-3.24). Contributing on social media was correlated with recent HIV testing (aOR 2.10, 95% CI 1.40-3.16), but neither observing (aOR 0.66, 95% CI 0.38-1.15) nor endorsing (aOR 1.29, 95% CI 0.88-1.90) were correlated. CONCLUSIONS:Our data suggest that social media use, particularly on multifunctional platforms such as WeChat and with contributing behaviors, is correlated with HIV testing among MSM in China. Campaigns that promote active participant contribution on social media beyond passive observation and endorsement of promotional materials are needed. This study has implications for the design and implementation of social media interventions to promote HIV testing.

Project description:PurposeThe purpose of this study is to analyze the Instagram practices of current orthopaedic surgeons and the components associated with highest reach and interactions.MethodsThe top 25 orthopaedic surgeon Instagram profiles using the hashtag #ortho were ranked by the number of followers. Account information regarding followers, posts, engagement percentage, average likes, average comments, average video view, average video likes, average video comments, and estimated cost per post was recorded using social media marketing tools. An analysis of each Instagram profiles' top 10 posts, based on number of likes, was conducted. A coding framework was developed to categorized posting strategies utilized.ResultsTwenty-five Instagram accounts and 250 Instagram posts were included in the analysis. Accounts with increased engagement rating had a significantly greater number of likes and video views. When examining post characteristics that influenced the number of likes a post generated, posts that elicited negative emotions received 52.6% and 70.7% more likes than positive emotions (P = .04) and neutral emotions (P < .01), respectively. Upon assessment of posting characteristics that influenced the number of comments a post generated, promotional posts were shown to have 43.7% fewer comments than nonpromotional posts (P = .02). When evaluating factors that influenced total number of interactions a post generated, it was found that posts that asked questions generated 26.4% more interactions (P < .01) than those that do not.ConclusionsThe present investigation found that the most effective strategies to generate more interactions on Instagram are those that elicit emotional responses and provoke viewer engagement by asking questions and directing actions. Additionally, it was found that promotional content was not well received by viewers.Clinical relevanceOrthopaedic surgeons have an opportunity to connect with colleagues, patients, and interested viewers through social media platforms in order to enhance their practice, disseminate educational content, and contribute to the social media presence of orthopaedic surgery.

Project description:BACKGROUND:Affective states, including sentiment and emotion, are critical determinants of health. However, few studies among men who have sex with men (MSM) have examined sentiment and emotion specifically using real-time social media technologies. Moreover, the explorations on their associations with sexual and health status among MSM are limited. OBJECTIVE:This study aimed to understand and examine the associations of affective states with sexual behaviors and health status among MSM using public data from the Blued (Blued International Inc) app. METHODS:A total of 843,745 public postings of 377,610 MSM users located in Guangdong were saved from the Blued app by automatic screen capture. Positive affect, negative affect, sexual behaviors, and health status were measured using the Simplified Chinese Linguistic Inquiry and Word Count. Emotions, including joy, sadness, anger, fear, and disgust, were measured using the Weibo Basic Mood Lexicon. A positive sentiment score and a positive emotion score were also calculated. Univariate and multivariate linear regression models on the basis of a permutation test were used to assess the associations of affective states with sexual behaviors and health status. RESULTS:A total of 5871 active MSM users and their 477,374 postings were finally selected. Both positive affect and positive emotions (eg, joy) peaked between 7 AM and 9 AM. Negative affect and negative emotions (eg, sadness and disgust) peaked between 2 AM and 4 AM. During that time, 25.1% (97/387) of negative postings were related to health and 13.4% (52/387) of negative postings were related to seeking social support. A multivariate analysis showed that the MSM who were more likely to post sexual behaviors were more likely to express positive affect (beta=0.3107; P<.001) and positive emotions (joy: beta=0.027; P<.001), as well as negative emotions (sadness: beta=0.0443; P<.001 and disgust: beta=0.0256; P<.001). They also had a higher positive sentiment score (beta=0.2947; P<.001) and a higher positive emotion score (beta=0.1612; P<.001). The MSM who were more likely to post their health status were more likely to express negative affect (beta=0.8088; P<.001) and negative emotions, including sadness (beta=0.0705; P<.001), anger (beta=0.0058; P<.001), fear (beta=0.0052; P<.001), and disgust (beta=0.3065; P<.001), and less likely to express positive affect (beta=-0.0224; P=.02). In addition, they had a lower positive sentiment score (beta=-0.8306; P<.001) and a lower positive emotion score (beta=-0.3743; P<.001). CONCLUSIONS:The MSM social media community mainly expressed their positive affect in the early morning and negative affect after midnight. Positive affective states were associated with being sexually active, whereas negative affective states were associated with health problems, mostly about mental health. Our finding suggests the potential to deliver different health-related intervention strategies (eg, psychological counseling and safe sex promotion) on a social media app according to the affective states of MSM in real time.

Dataset Information

Depressive Emotion Detection and Behavior Analysis of Men Who Have Sex With Men via Social Media.

Publications

Depressive Emotion Detection and Behavior Analysis of Men Who Have Sex With Men <i>via</i> Social Media.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets