Two Sliding Window Control Based High Utility Pattern Mining
Swapnali Londhe, Rupesh G. Mahajan · 2018
Value of the item set is considered as a utility of that item set and find out the high utility of the item set is the aim of utility mining. Some time database parameters are considered to find out high utility pattern eg. Profit, cost etc. In proposed approach utility value of particular items used for utility mining. In day to day life high utility pattern mining play important role in applications. It's current hot topic in today's research area. Different existing algorithms are present in this area. It first of all identify the candidate itemset's by using their utilities, and simultaneously identify the exact utility of that candidate pattern. Problem of using this algorithm is large number of candidate itemset's are generated. But after computing exact utility it's clear that most of the candidate having no high utility. For generating profitable product manufacturing plan it's very important to understanding the customer preferences in industrial area. Sliding Window based pattern mining approach which considering the quantity, quality and cost of each product for generating high profitable product set, which employed to find out high utility pattern. For establishing highly profitable manufacturing plan, which allow corporation to maximize its revenue, high utility pattern mining is important aspect. Large amount of stream data related to customer purchase behavior used for establishing manufacturing plan. Recent preference of the customers also helps in generating manufacturing plans. This paper contains Two Sliding Windows Pattern mining Algorithm (TSW) scans complete dataset And divide it in to two parts. Both parts scan simultaneously in order to improve the search space. The proposed algorithm (TSW) scans the text using two sliding windows, allowing multiple alignments in the searching process. Introducing A list structure and a novel algorithm for generating high utility pattern over large data, with the help of Sliding Window Control Method. This approach avoid the generation of candidate pattern. Due to reducing candidate pattern, algorithm not required large amount of memory space as well as computational resources for verifying candidate patterns. By considering all this parameters, it's an efficient approach.