Identification of Mental Disorders through Text Mining on Social Media

• Item type: Trans(1)
• 公開日: 2024-02-28
• タイトル: Identification of Mental Disorders through Text Mining on Social Media
• タイトル / 言語: en
• タイトル: Identification of Mental Disorders through Text Mining on Social Media
• 言語: eng
• キーワード / 主題Scheme: Other
• キーワード / 主題: [事例紹介論文] natural language understanding, machine learning, text mining, mental disorders, content analysis
• 資源タイプ / 資源タイプ識別子: http://purl.org/coar/resource_type/c_6501
• 資源タイプ: journal article
• 著者所属: Nagoya Institute of Technology
• 著者所属: Nagoya Institute of Technology
• 著者所属: Nagoya Institute of Technology
• 著者所属: Nagoya Institute of Technology
• 著者所属: Nagoya Institute of Technology
• 著者所属(英): en
• 著者所属(英): Nagoya Institute of Technology
• 著者所属(英): en
• 著者所属(英): Nagoya Institute of Technology
• 著者所属(英): en
• 著者所属(英): Nagoya Institute of Technology
• 著者所属(英): en
• 著者所属(英): Nagoya Institute of Technology
• 著者所属(英): en
• 著者所属(英): Nagoya Institute of Technology
• 著者名: Julien, Ghali; Kosuke, Shima; Koichi, Moriyama; Atsuko, Mutoh; Nobuhiro, Inuzuka
• 著者名(英): Julien, Ghali; Kosuke, Shima; Koichi, Moriyama; Atsuko, Mutoh; Nobuhiro, Inuzuka
• 論文抄録 / 内容記述タイプ: Other
• 論文抄録 / 内容記述: Mental disorders are a growing concern worldwide, early detection and accurate identification can lead to better treatment outcomes. In this paper, we explore the use of machine learning techniques to vectorize and identify mental disorders from textual data. We used a database of 700,000 Reddit posts published in six different subreddit associated with mental disorders: Anxiety, Borderline Personality Disorder, Bipolar, Depression, Mental Illness, and Schizophrenia. We used the TF-IDF vectorizer to represent the posts as numerical vectors and applied UMAP dimensionality reduction to visualize the data in two dimensions. The application of UMAP allowed us identify clusters and to use k-means clustering to identify groups of posts with similar content. We found that some posts, despite being posted in specific mental disorders, were presenting signs of depression or anxiety. Our results suggest that machine learning techniques can help identify clusters and relationships between mental disorders, potentially leading to improved diagnosis and treatment.
• 論文抄録(英) / 内容記述タイプ: Other
• 論文抄録(英) / 内容記述: Mental disorders are a growing concern worldwide, early detection and accurate identification can lead to better treatment outcomes. In this paper, we explore the use of machine learning techniques to vectorize and identify mental disorders from textual data. We used a database of 700,000 Reddit posts published in six different subreddit associated with mental disorders: Anxiety, Borderline Personality Disorder, Bipolar, Depression, Mental Illness, and Schizophrenia. We used the TF-IDF vectorizer to represent the posts as numerical vectors and applied UMAP dimensionality reduction to visualize the data in two dimensions. The application of UMAP allowed us identify clusters and to use k-means clustering to identify groups of posts with similar content. We found that some posts, despite being posted in specific mental disorders, were presenting signs of depression or anxiety. Our results suggest that machine learning techniques can help identify clusters and relationships between mental disorders, potentially leading to improved diagnosis and treatment.
• 書誌レコードID / 収録物識別子タイプ: NCID
• 書誌レコードID / 収録物識別子: AA11464803
• 書誌情報: 情報処理学会論文誌数理モデル化と応用（TOM）; 巻 17, 号 1, p. 29-35, 発行日 2024-02-28
• ISSN / 収録物識別子タイプ: ISSN
• ISSN / 収録物識別子: 1882-7780
• 出版者 / 言語: ja
• 出版者: 情報処理学会