- 學位論文
基於Instagram的圖像，文本和行為數據預測憂鬱傾向
Predicting Depression Tendency Based on Image, Text and Behavior Data from Instagram
摘要
憂鬱症是一種常見但嚴重的精神疾病。它被歸類為情緒障礙，這意味著它的特點是消極的思想和情緒。隨著網路科技越來越發達，越來越多的人發布他們的生活故事，並在社群媒體上表達及抒發自己的情感。經研讀諸多文獻發現，可以從社群媒體的巨量與多元內容萃取並引導出特定表徵和預測憂鬱症的方法。它已被研究人員廣泛用於研究心理健康問題。但是，現有的大部分研究都集中在社群媒體的文本數據上。很少有研究同時考慮文本和圖像數據。在這項研究中，我們的目標是通過分析用戶在Instagram上的發文。利用發文中的文本、圖像和行為來預測一個人憂鬱傾向。我們採用有效的數據收集機制來收集憂鬱症和非憂鬱症使用者。接下來，我們從圖像，文本和行為中提取三組特徵，以構建我們預測的深度學習模型。我們研究了利用社群媒體發文作為信號來理解憂鬱症的潛力。我們的實驗結果表明，該模型能夠識別具有憂鬱傾向的用戶，F-score為82.3%。它可以極大地促進早期篩檢和檢測憂鬱症的開發工具。
並列摘要
Depression is common but serious mental disorder. It is classified as a mood disorder, which means that it is characterized by negative thoughts and emotions. With the development of the Internet technology, more and more people post their life story and express their emotion on social media. Social media can provide a way to characterize and predict depression. It has been widely utilized by researchers to study mental health issue. However, most of the existing study focus on textual data from social media. Few studies considering both text and image data. In this study, we aim to predict one’s depression tendency by analyzing image, text and behavior of their posts on Instagram. We employ an effective data collection mechanism to collect depression and non-depressive user accounts. Next, we extract three sets of features from image, text and behavior to build our predictive deep learning model. We examine the potential for leveraging social media postings as a signal in understanding depression. Our experiment results demonstrate that the proposed model could recognize users who have depression tendency with an F-1 score of 82.3%. It could greatly contribute to the developing of tools for early screening and detection of depression.