An Innovative Approach for Association Rule Mining In Grocery Dataset Based On Non-Negative Matrix Factorization And Autoencoder

Abstract

Data mining with Deep Neural Networks becomes a very promising research area. In the course of our daily lives, enormous amounts of transactional data are produced. As a result, the transactional dataset is growing exponentially. The handling of data is a major issue. To handle transactional data, tools and strategies are therefore required. The tool that accelerates the decision-making process and enables unique knowledge management of the information found in transaction data is data mining. Many writers provided their methods to handle transactional data and discover new patterns. This study aims to provide a new algorithm for data mining that performs currently used algorithms in terms of speed and interest. The effectiveness of the DAENMF-ARM algorithm will be demonstrated by comparing it with the most popular and widely used data mining algorithms, namely Apriori ECLAT and FP-Growth, and establishing association rules. The organization of the paper is: the 1^st portion is the introduction, the 2^nd portion is the literature work, the 3^rd portion is the methodology, 4^th portion is the experimental setup, and last portion is conclusion of the paper.