Tendency Of Depression Disorders From Language Expression In Social Media

Nowadays, with a prevalent use of social media sites, around 40% of the world population expresses their feeling freely through personal space on online social media. Everyone might be familiar with the experience of scrolling through Facebook timelines, Twitter feed or Instagram and sees some mournful posts or photos but behind the quirky posts and highly edited pictures, these could be a sign of depression. Major depression disorder is a common and severe illness able to found at all walks of life. It causes feelings of sadness and directly affects both personal life and social life of those who are depressed. According to American Foundation for suicide prevention, over 50 percent of all people who died by committing suicide suffer from major depression; moreover, depression negatively impacts nearly 5-8 percent of Americans ages 18 and over in research of 2016.

As same as the situation happening in Thailand, Ministry of Public Health revealed that more that 1. 5 million Thai citizens now are depressed and 50 percent of the entire suicide cases in 2017 were depressed patients. Therefore, from statistic report, we can see that major depression is a huge problem the world is facing. Fortunately, this ailment is treatable if we can detect. There is a way to link the tentative depressed individuals to the posts in social media. For example, in the journal EPJ Data Science, found that individuals suffering from depression were more likely to post Instagram photos that were bluer, darker, and greyer. Besides, in the case of language expression, Clinical Psychological Science has unveiled that people with symptoms of depression significantly use more first-person singular pronouns such as ‘me’, ‘myself’ and ‘I’. Another classifier for depression diagnosis is those who are depressed use more negative emotion words such as ‘waste’ and ‘hurt’ than non-depressed individuals. As well as anxiety-related words, for instance, ‘afraid’, and ‘confused’. Hence, it seems social media can be used to diagnose depression and should be employed to diagnose widely.

Major depressive disorder is considered as one of the significant issues in Thailand and the world. Depression causes suffering, loss of study and work ability, worsening relationship and, worse, committing suicide. At the same time, the era of social media is becoming more and more common. Social media are tools that many people use to communicate and convey thoughts, emotions or daily stories. Online media, therefore, are a virtual diary provided by digital 6 content. With its popularity, it leads to the question that can we observe whether a person is depressed from online media habits? These days have witnessed for the research about language indicators of depression which increase the possibility to detect depressed people from posts on online personal space. Most studies use a comparison of people diagnosed with depressive disorder with general population and the results of most studies are evidence-based with more that 70% correct. Depressed people often post or share messages with negative emotions or words related to depression, for example, bored, lonely, anxious, sad, painful, depressed, worthless, nobody, stress, etc. It also has been found that depressed patients are more likely to post or use online media late in the day than people who usually post in daytime. This is partly due to insomnia, which is one of the symptoms of depression. Moreover, there are some depressed patient beginnings asking for advice about depression treatment or indirect question themselves about stress. Thus, it is important that to find the proper indicator language usage which can point out the patient of depressive group.

Wang, Zhang, Ji, Li, and Wu (2013) “A Depression Detection Model Based on Sentiment Analysis in Micro-blog Social Network” was released at 2013. This paper applies data mining to psychology area for detecting depressed users in social network services. Firstly, a sentiment analysis method is proposed utilizing vocabulary and man-made rules to calculate the depression inclination of each micro-blog. Secondly, a depression detection model is constructed based on the proposed method and 10 features of depressed users derived from psychological research. Then 180 users and 3 kinds of classifiers are used to verify the model, whose precisions are all around 80%. Also, the significance of each feature is analyzed. Lastly, an application is developed within the proposed model for mental health monitoring online.

However, this paper is based on Chinese language and micro-blog social networks. A model for detecting depressed users in social network based on sentiment analysis is proposed in this paper. The sentiment analysis method pays special attention to the characteristics of depression and Chinese micro-blog content, and ten features are applied in the depression detection model. The precisions obtained from training dataset 7 are all around 80%, and the significance of each feature in the model is also analysed for model simplification. An application in Sina Micro-blog is developed to test the polarity of micro-blog and the mental state of users, which has helped psychologists detect several potential depressed users in Sina Micro-blog. Although the depression detection model is proposal based on Chinese vocabulary, the basic idea of the frame especially the sentence structure pattern mining and principle micro-blog features related to depression could be explicitly extended to other language scenarios. In this work, the training data detected by psychologists are widely scattered in social network.

It is hard to analyse the relationship between them, so user interaction is paid little attention and simply three parameters are considered. However, homophile is manifest in the group of depression users, which means, the friends of the depression are more likely to be depressive, and different kinds of interactions indicate different results. Thus the influence of ties between users is contemplated to be studied in the future and a deeper understanding about depression in SNS will be provided.