ADAPTIVE-CLUSTERING BASED METHOD TO ESTIMATE NULL VALUES IN RELATIONAL DATABASES

Data preprocessing is an essential step of knowledge discovery. Data preprocessing comprises data cleaning, data integration, data transformation, data reduction and data discretization. Estimating null values is a task of data cleaning. Null values in a database are significant sources of poor data quality. Therefore, the appropriate handling of null values is an important task of data preprocessing in relational databases. We propose a new method that uses adaptive learning techniques, based on clustering, to resolve the issue of null values in relational database systems. This study uses clustering algorithms to group data and calculates the degree of influence between independent attributes (variables) and the dependent attribute through an adaptive learning method (the best adaptive parameter can be obtained by the minimum average error rate). Three databases (a human resource database, Waugh's database and a government salary study database) were selected as the experimental data to compare the mean absolute error rate (MAER) of the proposed algorithm with the other methods. The results demonstrate that the proposed method outperforms other methods.