ESTIMATING NULL VALUES IN RELATIONAL DATABASE SYSTEMS HAVING NEGATIVE DEPENDENCY RELATIONSHIPS BETWEEN ATTRIBUTES.

In recent years, some methods have been proposed to estimate the null values in relational database systems (Chang and Chen 2006; Chen and Chang 2008; Chen and Hsiao 2005; Chen and Huang 2003, 2008; Chen and Lee 2005; Cheng and Wang 2006). In this paper, we present a new method for estimating the null values in relational database systems having negative dependency relationships between attributes. First, we apply the automatic clustering algorithm we presented in Chang and Chen (2006) for clustering the tuples in the relational database. Based on the clustering results and multiple regression techniques (Kvanli et al. 1986), we present a new method for estimate null values in relational database systems having negative dependency relationships between attributes, where the “Benz secondhand car database” (Chen and Huang 2008) is used for the experiment. The experimental results show that the proposed method gets a higher average estimated accuracy rate than Chen and Huang's method (2003, 2008) for estimating null values in relational database systems having negative dependency relationships between attributes. [ABSTRACT FROM AUTHOR]