Using Text Mining to Explore The Influence of Food on The Death Risk of Adult

• Main Authors: Yen Wei Liu, 劉彥煒
• Other Authors: Ya-Han Hu
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/4bknp9

碩士 === 國立中正大學 === 資訊管理系醫療資訊管理研究所 === 107 === With the advancement of technology, human beings have a variety way of cooking and preserving food, such as microwave food, canned food, vacuum food, ready-to-eat and food. People have more choices of food. Because of working, people don’t have meals in right time. Microwave food and ready-to-eat food are the best choice for people. However, these kind of foods increase risk of death and illness. Past studies have confirmed that some foods have a high risk of death, such as red meat, high-cholesterol foods and processed foods. Due to the advance of technology, some undiscovered potential dangerous foods haven’t found the impact on health and death by past studies. The purpose of this study is to identify potentially dangerous foods. We use the US NHANES database to find out the association between food and death from this large database by using association rules. Then, we collect the literature from the PubMed website as verification for an interesting rule. In addition, we use automated distinguish article results. We ask expert to evaluate whether the article results match the association rules that we found. We also evaluate the accuracy of the results of this automated distinguishing. After the experiment analysis, the study obtained 10 interesting rules. We found potential dangerous foods which didn’t find in past studies, such as frozen food, ready-to-eat meal. They can be explored by future researchers to understand whether their risk to health and death is high or low. This study invented a method for automatically distinguish articles, which automatically distinguished the accuracy of the article to meet the expert evaluation.