The study of value-added process in database: the case of agriculture database
Date Issued
2004
Date
2004
Author(s)
Chiang, Chih-Ming
DOI
zh-TW
Abstract
According to the limitation and insufficiency of planning for constructing the database system at the beginning, the problem, lack of data (variables), arises when executing data mining. How to add every kind of data (variables) to database?” , therefore, is the important issue to the researchers. It is possible to continue the work of data mining if acquiring another useful and appropriate auxiliary information, such as other database or secondary data from other unit. However, the auxiliary information cannot be used and inserted in database immediately because of their structure. The purposes of this study are to deal with the two kinds of problem. One is “ functional mapping from the other database”, the other is “ value-added data from secondary data”.
First we propose a process to try to combine 2 similar databases cannot be merged directly for functional mapping from the other database, named “DISBME”. Second we will discuss that how to deal with the kind of percentage data for value-added data from secondary data. Then the practices of real database will be conducted. The 2 databases are “2000 agriculture and husbandry census” and “2001 agriculture and husbandry sampling survey”.
The results show that both are satisfied and don’t cause much bias. We add the useful variables in the main database successfully. It means, through above process, the main database can achieve reliable data from each kind and issue of auxiliary information. It also resolves the problem of insufficiency of data and increases the value of the database.
Subjects
資料採礦
加值預測模式
資料加值
函數映射
資料短缺
預測
predictive model for data added
data mining
value-added
data-added
data insufficiency
functional mapping
Type
thesis