A data mining approach to assessing the extent of damage of missing values in survey. (22nd October 2007)

• Record Type: Journal Article
• Authors: Wang, Hai; Wang, Shouhong
• Is Part Of: International journal of business intelligence and data mining. Volume 2:Number 3(2007)
• Page Start: 249
• Page End: 260
• Publication Date: 2007-10-22
• Subjects: business intelligence -- data mining -- knowledge discovery -- data quality -- incomplete data -- missing values -- survey data -- classification -- fuzzy sets -- fuzzy observations; 006.312
• Languages: English
• ISSNs: 1743-8187

Abstract

Survey data are often incomplete. Incomplete data are often mistreated and damages of missing values in survey are often overlooked in data mining. This study proposes a classification-based data mining approach to assessing the extent of damage of missing values in survey. Using this approach, an incomplete observation is translated into fuzzy observations. These fuzzy observations are used to test the classifier that has been trained by the complete data set of the survey. The test results provide a base for discovering knowledge about the implication of missing data and the quality of the survey.