SenseMood: Depression Detection on Social Media

Lin, Chenhao; Hu, Pengwei; Su, Hui; Li, Shaochun; Mei, Jing; Zhou, Jie; Leung, Henry W.C.

doi:10.1145/3372278.3391932

Cited by 92 publications

(38 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Research has shown how a neural network can be designed to detect depression with limited data and without any exhaustive feature engineering [31], presenting a neural network architecture that optimizes word embeddings. SenseMood [32] is applying a CNN-based classifier and Google's Bert model [33] on posted images and tweets from users with or without depression, combining visual and textual features. They are using a dataset previously presented in research [34], which contains a set of users with anchor tweets matching the strict pattern that they have been diagnosed depression.…”

Section: Natural Language Processing For Burnout/depression Detectionmentioning

confidence: 99%

BurnoutWords - Detecting Burnout for a Clinical Setting

Nath¹,

Kurpicz-Briki²

2021

Machine Learning Techniques and Data Science

View full text Add to dashboard Cite

Burnout, a syndrome conceptualized as resulting from major workplace stress that has not been successfully managed, is a major problem of today's society, in particular in crisis times such as a global pandemic situation. Burnout detection is hard, because the symptoms often overlap with other diseases and syndromes. Typical clinical approaches are using inventories to assess burnout for their patients, even though free-text approaches are considered promising. In research of natural language processing (NLP) applied to mental health, often data from social media is used and not real patient data, which leads to some limitations for the application in clinical use cases. In this paper, we fill the gap and provide a dataset using extracts from interviews with burnout patients containing 216 records. We train a support vector machine (SVM) classifier to detect burnout in text snippets with an accuracy of around 80%, which is clearly higher than the random baseline of our setup. This provides the foundation for a next generation of clinical methods based on NLP.

show abstract

Section: Natural Language Processing For Burnout/depression Detectionmentioning

confidence: 99%

BurnoutWords - Detecting Burnout for a Clinical Setting

Nath¹,

Kurpicz-Briki²

2021

Machine Learning Techniques and Data Science

View full text Add to dashboard Cite

show abstract

“…At present, the research platforms for online depression detection are mainly focused on Twitter and Facebook (Islam et al, 2018 ; Alsagri and Ykhlef, 2020 ; Lin et al, 2020 ), and many researchers are trying to use different natural language processing methods to determine the characteristics of their contributions to build an accurate detection model. Due to the heterogeneity among social media, multiple features ranging from text and language to user-based and metadata can be extracted.…”

Section: Introductionmentioning

confidence: 99%

A Hybrid Feature Selection and Ensemble Approach to Identify Depressed Users in Online Social Media

Liu

Shi

2022

Front. Psychol.

View full text Add to dashboard Cite

Depression has become one of the most common mental illnesses, and the widespread use of social media provides new ideas for detecting various mental illnesses. The purpose of this study is to use machine learning technology to detect users of depressive patients based on user-shared content and posting behaviors in social media. At present, the existing research mostly uses a single detection method, and the unbalanced class distribution often leads to a low recognition rate. In addition, a large number of irrelevant or redundant features in high-dimensional data sets interfere with the accuracy of recognition. To solve this problem, this paper proposes a hybrid feature selection and stacking ensemble strategy for depression user detection. First, recursive elimination method and extremely randomized trees method are used to calculate feature importance and mutual information value, calculate feature weight vector, and select the optimal feature subset according to the feature weight. Second, naive bayes, k-nearest neighbor, regularized logistic regression and support vector machine are used as base learners, and a simple logistic regression algorithm is used as a combination strategy to build a stacking model. Experimental results show that compared with other machine learning algorithms, the proposed hybrid method, which integrates feature selection and ensemble, has a higher accuracy of 90.27% in identifying online patients. We believe this study will help develop new methods to identify depressed people in social networks, providing guidance for future research.

show abstract

“…Depressive disorders have been a recent research concern in the field of Data Science [18], especially after the discovery that data mining approaches can assist in the diagnostic identification and treatment of depression [19], [20]. In a previous study [21], we provided an extensive discussion about recognizing depressive mood disorders in social networks using machine learning and sentiment analysis.…”

Section: Introductionmentioning

confidence: 99%

“…In contrast, social support established in social networks was associated with an increased risk of developing the disorder (effect size, 20%). Considering the exponential increase of users on social networks in recent years [20], evaluating the network structure and its impact on users' mental health has been of interest to several researchers in the area.…”

Section: Introductionmentioning

confidence: 99%

Modeling and Assessing the Temporal Behavior of Emotional and Depressive User Interactions on Social Networks

et al. 2021

View full text Add to dashboard Cite

The way users interact on social media can indicate their well-being. When depressed, people's feelings tend to be more evident, affecting how users interact and demonstrate their feelings on social media. This paper presents a new approach for the temporal assessment of emotional behavior and interaction among depressed users on social networks. We start by modeling user interactions using complex networks, grouping users through time using the Clauset-Newman-Moore greedy modularity maximization. We evaluate the built networks using metrics such as assortativity, density, clustering, diameter, and shortest path length, closeness, and coverage. Then, we propose EMUS, a method for establishing an emotional user score based on the extraction of emotional features in texts of posts and comments. To extract emotional features, we combine the use of the Empath framework and VADER lexicon. Finally, based on the standard deviation among users, we establish a metric for assessing mood levels. We evaluated users for 33 days, and the results show a sequence of mixed emotional behaviors with high correlations between the number of active users in the network communities, and the form and quality of interactions. The developed approach can be further applied to other database graphs, for different sequential patterns analysis and text-mining contexts.

show abstract

SenseMood: Depression Detection on Social Media

Cited by 92 publications

References 16 publications

BurnoutWords - Detecting Burnout for a Clinical Setting

BurnoutWords - Detecting Burnout for a Clinical Setting

A Hybrid Feature Selection and Ensemble Approach to Identify Depressed Users in Online Social Media

Modeling and Assessing the Temporal Behavior of Emotional and Depressive User Interactions on Social Networks

Contact Info

Product

Resources

About