Temporal Contrast Sets Mining

Discovering discriminating features’ values that distinguish different data groups is vital for understanding the unique characteristics that define each group. These distinct features can be utilized in various applications, such as classification and data mining. One effective technique for this task is Contrast Sets Mining, which identifies sets of attribute-value pairs that differentiate groups. However, existing contrast sets mining techniques overlook the temporal dimension of data, which is critical in certain applications like disease identification based on ordered symptoms. To address this limitation, this work introduces a novel approach called Temporal Contrast Sets Mining, which leverages sequential association rules to capture the temporal aspect of the data. The proposed model is evaluated using a real dataset of students’ academic performance. The results are discussed, providing valuable insights for educators and students to gain a better understanding of the significant sequences that influence students’ performance. By incorporating the temporal dimension into contrast sets mining, this research contributes to the advancement of techniques for analyzing ordered data, thereby enhancing the applicability of contrast sets in various domains.