Communications of the Association for Information Systems


Data mining is a useful analytic method and has been increasingly used by organizations to gain insights from large-scale data. Prior studies of data mining have focused on developing automatic data mining models that belong to first-order data mining. Recently, researchers have called for more study of the second-order data mining process. Second-order data mining process is an important step to convert data mining results into intelligent knowledge, i.e., actionable knowledge. Specifically, second-order data mining refers to the post-stage of data mining projects in which humans collectively make judgments on data mining models’ performance. Understanding the second-order data mining process is valuable in addressing how data mining can be used best by organizations in order to achieve competitive advantages. Drawing on the theory of habitual domains, this study developed a conceptual model for understanding the impact of human cognition characteristics on second-order data mining. Results from a field survey study showed significant correlations between habitual domain characteristics, such as educational level and prior experience with data mining, and human judgments on classifiers’ performance.