應用詞向量於語言樣式探勘之研究

Mining Language Patterns Using Word Embeddings

負面生活事件是引發抑鬱情緒的重要原因，比如家庭成員的去世、與伴侶的爭吵、被老闆開除或被老師責備等。主語和負面生活事件的組合稱為負面生活事件語言樣式。能否通過自動準確識別這些負面的生活事件來理解那些有抑鬱傾向的網絡文本對建立有效實用的精神病學網絡服務是至關重要的。如問題（父母：離婚::男友：？），本文使用分佈式詞向量模型來挖掘負面生活事件語言樣式，實驗結果MRR值達0.71。

Negative life event is an important reason of causing depression, such as the death of family members, quarreled with the spouse, fired by the boss or blamed by the teacher. The combination of a subject and a negative life event is called a language pattern of negative life events. Therefore, whether it can understand those web texts with depression trend by identifying these negative life event language patterns automatically and accurately, which is important to establish effective and practical psychiatric network services. Such as the question (parent: divorce:: boyfriend:?), this study applied the distributional word embeddings to mine the language patterns of negative life events. The experimental results showed that the MRR index reached 0.71.