Frequent Itemset Generation using Clustering

Data mining examines pre-existing large databases to generate new information. For large databases, research is needed to improve mining performance and accuracy, therefore, much of the current focus on association rule mining lies in new mining theories, algorithms, and mining algorithms. improve old methods. Association rules mining is a function of the data mining research domain and arouses many researchers' interest to design a highly efficient algorithm to mine association rules from transaction databases. With the increasing size of the database, we have a large amount of data, but unfortunately we cannot use raw data in our daily decisions/inferences. We desperately need specific information. This information is in most cases in the gathered data, but the extraction of it is a very time and resources consuming operation. In the single machine environment, the problems of Apriori and FP-Growth algorithm in large-scale data association rules mining are high memory consumption, low computing performance, poor scalability and reliability, choosing appropriate minimum support and so on. The purpose of this project is to introduce an improved FP-growth algorithm that will help to resolve the bottleneck problems of the traditional algorithm and has more efficiency than the original one. Therefore, we put forward a new implementation method which depends on a cluster based approach for mining frequent item sets to generate association rules and is verified by using real datasets. Experimental results show that the proposed algorithm was able to reduce the size of constructed trees and the execution time was significantly less than that of the conventional FP-Growth algorithm.