Award Date

May 2017

Degree Type

Thesis

Degree Name

Master of Science in Computer Science

Department

Computer Science

First Committee Member

Justin Zhan

Second Committee Member

Laxmi Gewali

Third Committee Member

Fatma Nasoz

Fourth Committee Member

Ge Lin Kan

Number of Pages

55

Abstract

High utility itemset mining is an important data mining problem which considers profit factors besides quantity from the transactional database. It helps find the most valuable products/items that are difficult to track using only the frequent data mining set. An item that has a high-profit value might be rare in the transactional database despite its tremendous importance. While there are many existing algorithms which generate comparatively large candidate sets while finding high utility itemsets, the major focus is to reduce the computational time significantly with the introduction of pruning strategies. Another aspect of high utility itemset mining is to compute the large dataset. There are very few algorithms that can handle a large dataset to find high utility itemset mining in a parallel (distributed) system.

In this thesis, there are two proposed methods: 1) High utility itemset mining using pruning strategies approach (HUI-PR) and 2) Parallel EFIM (EFIM-Par). In the method I, the proposed algorithm constructs the candidate sets in the form of a tree structure, which traverses the itemsets with High Transaction-Weighted Utility (HTWUIs). It uses a pruning strategies to reduce the computational time by refraining the visit to unnecessary nodes of an itemset to reduce the search space. It significantly minimizes the transaction database generated on each node. In the method II, the distributed approach is proposed dividing the search space among different worker nodes to compute high utility itemsets which are aggregated to find the result. The experimental results for both methods show that they significantly improve the execution time for computing the high utility itemsets.

Keywords

Data mining; itemset mining; parallel computing; spark

Disciplines

Computer Sciences

Language

English


Share

COinS