本論文計一冊，約四萬字，內分六章二十節。 第一章為緒論，說明本論文研究目的和範圍，敘述研究方法與資料來源，並提出本研究之限制。 第二章對判斷抽樣與系統抽樣的意義、型態及技術加以介紹。 第三章介紹國外有關統計抽樣之實證研究，以及實務上之應用情形，並就判斷抽樣與統計抽樣的優、缺點加以比較。 第四章係根據文獻資料及筆者個人經驗，探討可能影響審計人員抽樣方法的各種因素。 第五章則將實證方法搜集而得的資料。利用因素分析方法建立可能影響審計人員抽樣方法因素之構面，並就每個構面，分析台灣審計人員對這些因素的看法。 最後一章對本研究之結果做一結論，並根據本研究之缺點，建議將來之研究方向。

在本論文中，我們參考國外隨機撥號的方法，作了三個階段的測試。第一個階段為後一碼隨機以及後四碼隨機的測試。第二階段為了研究局碼，有二分之一局碼的選取、四分之一局碼選取、後一碼隨機以及後四碼隨機四種抽樣方法。第三階段主要針對後四碼隨機的方法進行改良，產生後三碼隨機的方法。經過實際的測試比較，在局碼選取上二分之一局碼的選取可行，而在尾數的處理上，後三碼隨機的方法較佳。

In the wake of growing database that has already become the trend of today’s business environment within the foreseeable future, reviewing quality information from mountains of data residing on corporations or organizations’ network such as sales figures, manufacturing statistics, financial data and experimental data is clearly costly, time consuming and definitely ineffective approach. Therefore we would need a sound and effective method in obtaining only portions of the data that are representative to the population and which allow us to build the reliable model based upon the sampled data. However, sometimes we have a situation where the database is of limited in size, under such circumstance, we initiate the idea which is relatively new to adding the attributes or values into the database to enhance the quality of the data Follow through such a procedure; it is obvious that implementing a good sampling method is an important groundwork leading us to reach final destination that is obtaining a reliable predictive model. And this is our research goal that is to get an effective and representative value-added sample of by means of sampling method for building an accuracy predictive model. The concept is pretty straightforward that is if we want to get good predictive samples then we need the correct sampling methods. The sampling methods under study are simple random sample, system sample, stratified sample and uniform design. The models used are the C5.0, logistic regression, and neural network for categorical predictive variable and stepwise regression for continuous predictive variable. The results are discussed in the conclusion section. Keywords: Database、Data Mining、Sampling、Value-added database

有賴電腦的輔助，企業或組織內部所存放的資料量愈來愈多，加速資料量擴大的速度。但是大量的資料帶來的未必是大量的知識，即使擁有功能強大的資料庫系統，倘若不對資料作有意義的分析與推論，再大的資料庫也只是存放資料的空間。過去企業或組織只把資料庫當作查詢系統，並不知道可以藉由資料庫獲取有價值的資訊，而其中資料庫的內容完整與否更是重要。由於企業所擁有的資料庫未必健全，雖然擁有龐大資料庫，但是其中資訊未必足夠。我們認為利用資料庫加值方法：插補方法、抽樣方法、模型評估等步驟，以達到擴充資訊的目的，應該可以在不改變原始資料結構之下增加資料庫訊息。 本研究主要在比較不同階段的資料經過加值動作後，是否還能與原始資料結構一致。研究架構大致分成三個主要流程，包括迴歸模型、羅吉斯迴歸模型與決策樹C5.0。經過不同階段的資料加值後，我們所獲得的結論為在迴歸模型為主要流程之下，利用迴歸為主的插補方法可以使加值後的資料庫較貼近原始資料，若想進一步採用抽樣方法縮減資料量，系統抽樣所獲得的結果會比利用簡單隨機抽樣來的好。而在決策樹C5.0的主要流程下，以類神經演算法作為插補的主要方法，在提增資訊量的同時，也使插補後的資料更接近原始資料。關於羅吉斯迴歸模型，由於間斷型變數的類別比例差異過大，致使此流程無法達到有效結論。 經由實證分析可以瞭解不同的配模方式，表現較佳的資料庫加值技術也不盡相同，但是與未插補的資料庫相比較，利用資料庫加值技術的確可以增加資訊量，使加值後的虛擬資料庫更貼近原始資料結構。 / With the fast pace of advancement in computer technology, computers have the capacity to store huge amount of data. The abundance of the data, without its proper treatment, does not necessary mean having valuable information on hand. As such, a large database system can merely serve as ways of accessing and storing. Keeping this in mind, we would like to focus on the integrity of the database. We adapt the methods where the missing values are imputed and added while leaving the data structure unmodified. The interest of this paper is to find out when the data are post value added using three different imputation methods, namely regression analysis, logistic regression analysis and C5.0 decision tree, which of the methods could provide the most consistent and resemblance value-added database to the original one. The results this paper has obtained are as the followings. The regression method, after imputation of the added value, produced the closer database structure to the original one. And in the case of having large amount of data where the smaller size of data is desired, then the systematic sampling provides a better outcome than the simple random sampling. The C5.0 decision tree method provides similar result as with the regression method. Finally with respect to the logistic regression analysis, the ratio of each class in the discrete variables is out of proportion, thereby making it difficult to make a reasonable conclusion. After going through the above studies, we have found that although the results from three different methods give slight different outcomes, one thing stands out and that is using the technique of value-added database could actually improve the authentic of the original database.