Applying Data Mining to Explore the Influencing Factors of Postpartum Depression

• Main Authors: YOU,YA-HAN, 游雅涵
• Other Authors: LIN,I-CHUN
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/31355279551351774314

碩士 === 國立雲林科技大學 === 工業工程與管理系 === 104 === 　　After delivery, hormones in a woman's body change, which affect their physiology and psychology. Postpartum depression is one of the most common psychological disorders after childbirth. Postpartum depression may affect woman's ability to care for herself and her family, and it may also lead to harm herself or her infant. Therefore, this study used data mining technique to construct a prediction model for postpartum depression, and explored the influencing factors of postpartum depression. 　　This study selected women who were delivered during 2006 -2010 periods from Taiwan National Health Insurance Research Database (NHIRD). The following 26 variables were used to construct the prediction model for postpartum depression by Decision Trees, Back-propagation Neural Network (BPN), and Support Vector Machine (SVM): age, delivery mode, delivery season, low income, catastrophic illness, 16 types of prenatal comorbidity, and 5 types of postpartum complications. 　　The results showed that, the performance of decision tree was superior to BPN and SVM. The main influencing factors of postpartum depression were antepartum depression, psychosis, delivery season, age, delivery mode and urinary tract infection. In addition, the variable of low income was also one of the influencing factors of postpartum depression. In the future, it could be discussing in depth.