Partial Weighted Count Tree for Discovery of Rare and Frequent Itemsets

Shwetha Rai, M. Geetha*, Preetham Kumar, B. Giridhar

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

1 Citation (Scopus)

Abstract

Time and space utilization for discovering interesting patterns from a database plays an important role in analyzing information for major sectors like education, medicine, and e-business. Association rule mining (ARM) technique is used to discover associations among the patterns from large volumes of data. In most ARM algorithms, rare and frequent itemsets discovery is optimized by mining pruned databases stored in the main memory. However, in this case, any change in requirements would necessitate re-scanning of the database. Weighted count tree (WC-Tree), and Single scan pattern tree (SSP-Tree) store the database in the main memory without pruning. WC-Tree stores the entire transaction as a node in the tree. However, if the weight is large, the actual information may be lost due to the precision error. In the current work, an efficient data structure, Partial weighted count tree (PWC-Tree), is proposed to store the database as a complete and compact structure in the main memory without losing the information. The work revealed that PWC-Tree construction is in O(n2) for n transactions in the database. The experimental results show that, for a large dataset, the PWC-Tree is time as well as space-efficient when compared with WC-Tree and SSP-Tree.

Original languageEnglish
JournalEngineered Science
Volume20
DOIs
Publication statusPublished - 01-12-2022

All Science Journal Classification (ASJC) codes

  • Chemistry (miscellaneous)
  • General Materials Science
  • Energy Engineering and Power Technology
  • General Engineering
  • Physical and Theoretical Chemistry
  • Artificial Intelligence
  • Applied Mathematics

Fingerprint

Dive into the research topics of 'Partial Weighted Count Tree for Discovery of Rare and Frequent Itemsets'. Together they form a unique fingerprint.

Cite this