用於社交網路上精神狀態偵測與描述之系統性方法

Systematical Approach for Detection and Description of Mental Status on Social Network

陳政翰(Zheng-Han Tan)

指導教授 ： 張玉山

國立臺北大學/電機資訊學院/資訊工程學系/碩士(2015年)

摘要

由於人們現在的生活步伐快速且忙碌，相對所承受的壓力也逐漸倍增，進而患上憂鬱症的人數也持續上升中。由於憂鬱症是一種無形的精神疾病，加上人們對於憂鬱症並不了解，所以人們往往在不知情的情況下患上憂鬱症，最終導致病情加重而傷害自己或家人。
隨著現今科技的進步與社交網路平台的蓬勃發展，人類對於分享內心的想法和意見已不再那麼的內斂。因此我們可以在如：臉書、微博等的網路社交平台上看到許多人分享心情記事、日記或是討論感興趣的議題等。
在此，我們從社交網路平台收集心情記事或日記，透過MENS (Mental Status Evaluator)系統去計算當事人的精神狀況以及他的嚴重程度為何。在系統中，我們使用監督式隱含狄利克雷分佈（SLDA, Supervised Latent Dirichlet Allocation）找出憂鬱症病患的特征并判斷是否有憂鬱症，接著使用隱含狄利克雷分佈（LDA, Latent Dirichlet Allocation）找出當事人最近文章中的潛在主題和最常出現的情緒字眼，加上SentiWordNet計算每個情緒字眼的分數，最後找出此分數與threshold的最短距離以表示其嚴重的程度。

關鍵字

憂鬱症 ； 隱含狄利克雷分布 ； 監督式隱含狄利克雷分布

並列摘要

Nowadays, human lifestyles were fast and busy, and the pressure also becomes relatively heavier. This situation causes the number of patients suffer from depression become higher. Depression is a kind of mental illness that is hard to detect in early stage. Due to the lack of knowledge in depression, many people were suffering from depression unconsciously and caused tragedy when sickness became more severe.
Due to the emergence of social platforms, people tend to posting their diaries and feeling online for sharing with others. In this study, we aim to predict whether a user is getting depressed or not through his blog posted on the Internet. For this purpose, we collect diaries from social network platform for diagnosis through MENS (Mental Status Evaluator) to calculate the mental health system and his severity. In the system, we use SLDA (Supervised Latent Dirichlet Allocation) to identify the characteristics of patients with depression and to determine whether they are getting depression or not. Then, we use LDA (Latent Dirichlet Allocation) to identify the latent topics and top frequent words from recent diaries. Finally, we use SentiWordNet to calculate the score of each emotion words and find the shortest distance through each training threshold to represent the severity of mental status.

並列關鍵字

Latent Dirichlet Allocation (LDA) ； Depression ； Supervised Latent Dirichlet Allocation (SLDA) ； SentiWordNet

參考文獻

1. Yassine, Mohamed, and Hazem Hajj. "A framework for emotion mining from text in online social networks." Data Mining Workshops (ICDMW), 2010 IEEE International Conference on. IEEE, 2010.

2. Thelwall, Mike, et al. "Sentiment strength detection in short informal text." Journal of the American Society for Information Science and Technology 61.12 (2010): 2544-2558.

3. Bao, Shenghua, et al. "Mining social emotions from affective text." Knowledge and Data Engineering, IEEE Transactions on 24.9 (2012): 1658-1670.

5.Nguyen, Thin, et al. "Affective and content analysis of online depression communities." Affective Computing, IEEE Transactions on 5.3 (2014): 217-226.

7. Dredze, Glen Coppersmith Craig Harman Mark. "Measuring post traumatic stress disorder in Twitter." (2014).