Mining Strong Substitution Rules between Sets of Items in Large Database
| Autor: | Yi-Chun Chen, 陳奕鈞 |
|---|---|
| Rok vydání: | 2007 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 95 Association rules mining problem has been studying for several years, while few works discuss on substitution rules mining. However, substitution rules mining will also lead to valuable knowledge in market prediction. In this thesis, the problem of mining substitution rules is discussed and SSM algorithm is proposed to solve it. Differ to previous works on substitution rules mining, only the strong substitution rules will be reported in this work. The idea of SSM algorithm can be decomposed into two stages: (1) generate frequent closed patterns; (2) utilize negative association rules and the Pearson correlation coefficient to mine out strong substitution rules based on the frequent closed patterns. Moreover, to make the mining process more efficient, two lemmas are proposed to prune the redundant substitution rules. The experiment results show that the SSM algorithm offers more excellent performance and finding less substitution rules. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|