Mining Frequent Itemsets for Evolving Database Involving Insertion

Abstract

Mining frequent itemsets is one of the popular task in data mining. There are many applications like location-based services, sensor monitoring systems, and data integration in which the content of transaction is uncertain in nature. This initiates the requirements of uncertain data mining. The frequent itemsets mining in uncertain transaction databases semantically and computationally differs from techniques applied to standard certain databases. The goal of proposed model is to deal with the problem of extracting frequent itemsets from evolving databases using Possible World Semantics (PWS). As evolving databases contains exponential number of possible worlds mining process can be modeled as Poisson Binomial Distribution (PBD). In this proposed work apriori-based PFI mining algorithm and approximate incremental mining algorithm are developed. An approximate incremental mining algorithm can efficiently and accurately discover frequent itemsets. Also, focus is on the issue of maintaining mining results for uncertain databases.
DOI: 10.17762/ijritcc2321-8169.1506159