| DC 欄位 |
值 |
語言 |
| DC.contributor | 資訊工程學系 | zh_TW |
| DC.creator | 張立偉 | zh_TW |
| DC.creator | Li-Wei Chang | en_US |
| dc.date.accessioned | 2017-7-26T07:39:07Z | |
| dc.date.available | 2017-7-26T07:39:07Z | |
| dc.date.issued | 2017 | |
| dc.identifier.uri | http://ir.lib.ncu.edu.tw:88/thesis/view_etd.asp?URN=104522032 | |
| dc.contributor.department | 資訊工程學系 | zh_TW |
| DC.description | 國立中央大學 | zh_TW |
| DC.description | National Central University | en_US |
| dc.description.abstract | 隨著台灣人口高齡化和生活習慣的變化，癌症正成為一個重要的健康問題，因此對於癌症的預防和政策是至關重要的。至今還未曾有一個廣泛並同時討論多種疾病與所有癌症的關聯性在同一個人群資料中。本論文旨利用資料探勘的關聯式規則分析，以台灣人口健保資料庫為基底來分析所有疾病與癌症之間的關係。我們將所有疾病根據疾病類型分成 17 群、癌症分成 10 群，為了要更精準地分析他們之間的關係。將已分群的 SAS 資料，利用 SAS 程式碼和 C 語言轉換成 WEKA 的格式，藉由 WEKA 軟體進行關聯式的分析預測。分析後得到的結果，我們針對所有
疾病與特定癌症部位，利用現有發表過的文章來印證結果；並挑選出 6 組值得探討的疾病來更細節的比較現有文獻與我們的差異。雖然我們得到的結果跟現有文獻上的結果有些出入，但相似部分還是占大多數，而且也發現了一些沒有被文獻討論到的關係。本 研究完成了一個利用關聯式分析全面性的探索所有疾病與癌症之間的關係，並整理出癌症的相關文獻，讓大家更注意到台灣癌症預防的重要性。 | zh_TW |
| dc.description.abstract | Introduction: With the issue of cancer is being seriously, more and more cohort studies are concerned about the risk of developing cancer in a particular disease. In recent years, data mining is widely used in large data analysis which is an automatic or semiautomated way to explore and analyze large amounts of data. Therefore, the use of data mining techniques to comprehensively explore the association between all diseases and malignant neoplasms is worth to be concerned. The aim of this study is to use association rule to comprehensively explore the relationship between cancer and other diseases in the same Taiwan population database. Methods: Base on Taiwan population database NHIRD, we use LHID which is a sample of one million beneficiaries randomly sampled from NHIRD for a specific year from 2005 to 2013. After finding the malignant cancer patients above 30 years old with ICD-9-CM 140~209, we divide all diseases into 17 groups and all cancers into 10 groups. Then, we use WEKA to do our association rule to find the relevance of diseases and cancers. Results: After we get the relationship between diseases and cancer, we sort out the results according to the disease groups. In order to verify our results by current knowledges, we tried to compare our results with their difference. Then, we select six diseases worthy of discussion to more detailed comparison the difference between current literatures and our results. Discussions and Conclusions: This study succeeded comprehensive search for all the related diseases before cancers by using data mining method association rule in the same Taiwan population database. Although during verify, there are some difference between our results and previous works, the similar between them are in majority. And also sort out lots of cancer-related studies for cancer prevention. | en_US |
| DC.subject | 全民健保資料庫 | zh_TW |
| DC.subject | 廣泛搜索 | zh_TW |
| DC.subject | 癌症 | zh_TW |
| DC.subject | 資料探勘 | zh_TW |
| DC.subject | 關聯法則 | zh_TW |
| DC.subject | NHIRD | en_US |
| DC.subject | comprehensive explore | en_US |
| DC.subject | cancer | en_US |
| DC.subject | data mining | en_US |
| DC.subject | association rule | en_US |
| DC.title | 利用關聯分析全面性的搜索癌症關聯疾病 | zh_TW |
| dc.language.iso | zh-TW | zh-TW |
| DC.title | Using Association Rules to Comprehensively Search for the Relations between Malignant Neoplasm and Diseases | en_US |
| DC.type | 博碩士論文 | zh_TW |
| DC.type | thesis | en_US |
| DC.publisher | National Central University | en_US |