Dirty Data Impacts on Regression Models

Due to the negative influence of dirty data on the accuracy of regression models, the relation between the data quality and model results is able to be used in the selection of proper regression models and dirty data repairing strategies. Motivated by this, we develop an evaluation framework to measure the dirty data impacts on regression models. Based on the framework, we compare the impacts of missing, inconsistent, and conflicting data comprehensively. According to the evaluation observations, we suggest users how to select appropriate regression models and clean dirty data. Section 3.1 provides the research background of this chapter. Section 3.2 presents the methodology to measure the dirty data impacts on regression models and designs a generalized evaluation framework. Section 3.3 discusses the experimental observations of dirty data impacts on regression model results and gives guidelines of regression model selection and dirty data cleaning. Finally, we conclude this chapter in Sect. 3.4.