Making Decision Rules for the Discretize Data Attributes by using Association Algorithms
| Autor: | Yi-Ho Yang, 楊怡和 |
|---|---|
| Rok vydání: | 2005 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 93 In this paper we have presented a new algorithm of ICI-C (ICI-Classification) that improves on ICI (Incremental Combination Itemsets) algorithms; it successfully solves the problems of classification and correctly build the decision tree. ICI-C (ICI-Classification) algorithm utilizes the characteristic of ICI that set up the min-supports afterwards and is different from Apriori algorithm which spends a lot of searching time for adjusting min-support repeatedly for discovering association rules which is suitable for the classification. Because of setting up the min-support afterwards ICI algorithm saves a lot of searching time for discovering association rules and has higher performance and correctness. In this paper we use ICI-C algorithm and adjust the min-supports to find out the candidate itemsets by association rules and build the decision tree for classification. ICI-C algorithm can help administrator and policymaker to seek the portfolio of goods quickly, simple and correctly. To provide decision-makers with optimality when they mine information in large transaction in order to take out promotes goods for sale. The algorithm has been applied to UCI domains as well as to the COIL challenge data. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|