隨著癌症的議題越來越被重視，探討癌症風險的世代研究也逐漸地增加，因此對於進行流行病學研究的效 率也應該要被關注。至今還未曾有一個同時討論一種疾病與多種癌症關聯性的分析系統。本研究旨在建立 一個自動化分析系統，以台灣人口為基底來分析疾病及各癌症之間的關聯性。我們建立一個網頁來引導研 究員循序漸進的填寫欲分析的資料，完成後系統會檢查其完整性並將查詢資料傳送到系統核心。接著將查 詢條件轉換成相對應的SAS 程式碼，藉由SAS 統計軟體進行資料分析。分析結束後系統會將輸出的結果 轉換成常見的統計圖表並呈現於網頁上; 另外，我們將建製空氣品質資料庫，空氣品質資料來自環保署空 氣品質監測站，屆時可提供空氣汙染源在北中南及花東地區相關資料的查詢；最後，將利用此空氣品質資 料庫與健保資料庫，分析空氣汙染源與各癌症發生率的關係；建立分析系統後，在第一部分:疾病及各癌 症之間的關聯性部份，我們會用資料探勘的關聯法則(Association Rules)及挑選刊登在不同期刊上有關疾病 與癌症間關係的論文來驗證我們的分析結果。 ;With the issue of cancer is being taken seriously, more and more cohort studies are conducted to investigate the risk of developing cancer. Therefore the efficiency of research on the epidemiologic studies ought to be concerned. Currently, none of analysis systems have considered the correlations between a disease and all cancer types simultaneously. The aim of this study is to implement an automatic analysis system to analyze correlations between a disease and some of site-specific cancers as well as correlations between a pollutant and some cancer types in a large Taiwanese cohort. We build a website to lead researchers making a query step by step. After each step is performed, the system checks the integrity of the query and sends it to the system core to transform into corresponding SAS codes. Our system executes the SAS file by the statistical software until the end of analysis. The system core converts the SAS output files into result tables and displays on website of analysis system. Then, we construct an air quality database that all air quality data from Taiwan Air Quality Monitoring Network in Taiwan Environment Environmental Protection Administration. A user then can use our web-based input to query the amount of different air pollutants. For each air pollutant, our system will calculate seasonal distribution of aggregated daily each air pollutant concentrations of all 73 monitoring stations from a period. Finally, we further analyze correlations between a pollutant and some cancer types in a large Taiwanese cohort. In the first topic above, we verify the results generated by our system using the outcomes published in journals. In the future, we will use association rules to mine the correlations between a disease and some of site-specific cancers
