透過您的圖書館登入
IP:3.142.98.108
  • 期刊

高效率探勘高頻項目集之演算法

High Efficient Algorithms for Mining Frequent Itemsets

摘要


從交易資料庫中探勘高頻項目集,是資料探勘領域中最重要的研究問題之一。本論文以交易資料為探勘的資料來源,每一筆交易資料包含消費者曾經購買的產品項目,從兩方面探勘高頻項目集:一是修改CDAR(cluster-decomposition association rule)演算法對交易資料分群的方式,設計一個CDPL(clustering the database with the prefix item and the length of transaction data)演算法探勘高頻項目集。從實驗評估中顯示,CDPL 演算法的執行效率優於CDAR演算法探勘出高頻項目集;二是考量當有新增或刪除交易資料的情況,以CDPL演算法的探勘步驟為基礎,設計一個UCDPL(updating association rules with the CDPL algorithm)演算法更新高頻項目集。從效能實驗中顯示,當有新增或刪除交易資料時,UCDPL演算法可以很有效率更新高頻項目集。

關鍵字

資料探勘 高頻項目集 CDAR CDPL UCDPL

並列摘要


”Mining frequent itemsets” from the transaction database is one of the most popular problems in the field of data mining. This paper uses transaction data as the source of mining, and each transaction data contains a consumer ever bought product items. We present two high efficient algorithms for mining frequent itemsets. One is to modify the CDAR algorithm to cluster transaction data, and to present a CDPL (clustering the database with the prefix item and the length of transaction data) algorithm to mine frequent itemsets. The experiments show that the performance of the CDPL algorithm is faster than the CDAR algorithm. The other is to mine frequent itemsets when adding or deleting transaction data, and to present an UCDPL (updating frequent itemsets with the CDPL algorithm) algorithm to update frequent itemsets. The experiments show that the UCDPL algorithm is more efficient to update frequent itemsets when adding or deleting transaction data.

並列關鍵字

data mining frequent itemsets CDAR CDPL UCDPL

被引用紀錄


陳尔榮(2013)。應用資料探勘技術於用戶折價券推薦系統之研究~以某網路公司為例〔碩士論文,國立臺北大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0023-2101201312481800

延伸閱讀