| Abstract | Summary mining aims to find interesting summaries for a data set and to use data mining techniques to improve the functionality of Online Analytical Processing (OLAP) systems. In this paper, we propose an interactive summary mining approach, called GenSpace summary mining, to find the interesting summaries based on user expectations. In the mining process, to record the user's evolving knowledge, the system needs to update and propagate new expectations. In this paper, we propose a linear method for consistently and efficiently propagating user expectations in a GenSpace graph. For a GenSpace graph where uninteresting nodes can be marked by the user before the mining process, we propose a greedy algorithm to determine the propagation paths in a GenSpace subgraph that reduces the time cost subject to a fixed amount of space. |
|---|
