Interactive Data Quality Assistance - An Approach for Min(d)ing the Quality of Data

摘要

In this paper we introduce the concept of an interactive data quality system. Today one of the most challenging goals of data processing is to achieve and to maintain high data quality. Especially when the data is added manually by multiple users. The idea was originated out of the need to analyze the data set from a recognition assistance system. The system supports aerial image analysts in the task of the object recognition. Depending on which object features the aerial image analyst selects the solution set of the object types gets more precise. Especially in this field mechanisms that improve high data quality are important. Therefore we developed an interactive data quality system that helps experts generating potential rules through correlation that describe the whole data set. More precisely, we search for rules within the data set that are in general valid for a certain group of object types.