Application of Machine Learning in Database Management System for Consumer Pattern Prediction in Big Data

Abstract

The development of digital technology has driven the growth of data on a large scale, marking the era of big data with new challenges and opportunities in business decision-making. Conventional Database Management Systems (DBMS) have limitations in handling the complexity of big data, especially in unstructured data analysis and real-time prediction needs. This research aims to examine the application of machine learning (ML) algorithms directly in DBMS to improve predictive capabilities in analyzing consumer behavior patterns. The method used is Systematic Literature Review (SLR) of articles published between 2021-2024 from Scopus, IEEE, ScienceDirect, and SpringerLink databases. Of the 150 articles found, 10 articles were selected for further analysis based on inclusion criteria and topic relevance. The results show an increasing trend in global research, especially from China, the United States, and India. The most commonly used ML algorithms include Decision Tree, Random Forest, K-Means, and Neural Network (including LSTM and CNN), with DBMSs ranging from MySQL, PostgreSQL, to MongoDB and Hadoop. The integration of ML into DBMS is proven to improve efficiency and accuracy in consumer pattern prediction, supporting faster and more relevant decision-making. This research recommends further development in DBMS architecture that is adaptive to the needs of machine learning-based analytics.