Associative classification using patterns from nested granules
Thomas W.H. Lui, David Chiu · International Journal of Granular Computing Rough Sets and Intelligent Systems · 2010
To facilitate interpretation and consider the internal association relationships between values of a pattern used in associative classification, a new form of multiple value pattern known as nested high-order pattern (NHOP) is presented. Taking an associative pair as information granule, the pattern is formed as multiple levels of association events. The general form of high-order pattern (HOP), that NHOP is a subtype, is identified as variable outcomes extracted from a random N-tuple. The pattern is detected by statistical testing if the occurrence is significantly deviated from the expected according to a prior model or null hypothesis. In this paper, we propose a classification method (called C-NHOP) based on nested high-order patterns. The rationale is that complex association patterns reinforce the underlying meaningfulness in interpreting regularity, thus, can provide a better understanding of the data domain. In evaluating our method using 26 UCI machine learning benchmark datasets, the experiments show a highly competitive and interpretable result.