Interactive Pattern Mining Using Discriminant Sub-patterns as Dynamic Features

Recent years have seen a shift from pattern mining process that has users define constraints before-hand, and sift through the results afterwards, to an interactive one. This new framework depends on exploiting user feedback learn quality function for patterns. Existing approaches weakness in they use static pre-defined low-level features, attempt independent weights representing their importance user. As alternative, we propose work with more complex features are derived directly ranking imposed by Those used be aggregated help drive right direction. Experiments UCI datasets show using higher-complexity leads selection of patterns better aligned hidden while being competitively fast when compared state-of-the-art methods.