| Autor: |
Mong Li Lee, Kian Lee Tan, Wuwongse, Vilas, Gao Cong, Bin Cui, Yingxin Li, Zonghong Zhang |
| Zdroj: |
Database Systems for Advanced Applications (9783540333371); 2006, p171-186, 16p |
| Abstrakt: |
Frequent pattern mining is an important data mining problem with wide applications. The huge number of discovered frequent patterns pose great challenge for users to explore and understand them. It is desirable to accurately summarizing the set of frequent patterns into a small number of patterns or profiles so that users can easily explore them. In this paper, we employ a probability model to represent a set of frequent patterns and give two methods of estimating the support of a pattern from the model. Based on the model, we develop an approach to grouping a set of frequent patterns into k profiles and the support of frequent pattern can be estimated fairly accurately from a relative small number of profiles. Empirical studies show that our method can achieve compact and accurate summarization in real-life data and the support of frequent patterns can be restored much more accurately than the previous method. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Supplemental Index |
| Externí odkaz: |
|
