Authors:
Swathi Chundru

Addresses:
Department of Quality Control, Motivity Labs Inc., Irving, Texas, United States of America. swathichundru19@gmail.com

Abstract:

This paper presents the possibility of changing metadata management through AI and ML and proposes an approach to classifying and discovering metadata. This work takes a sample dataset with time-based impedance values ranging from 12.3 to 45.8 seconds and accuracy levels ranging from 0.87 to 0.95 to compare the performance of the various machine learning algorithms, such as decision trees, SVM, random forests, and CNN. It evaluates the performance of AI-based metadata discovery systems on different types of datasets, including healthcare, social media, and finance, by using Python as the core tool. The results found were that CNNs delivered the maximum accuracy of 0.95; however, they consumed more computation compared to others; simpler models like decision trees, which produced lesser accuracy, did so in lesser computation time. It is based on mixed bar-line graphs and impedance charts, describing a trade-off between speed and accuracy. Beyond that, it established the fact that AI systems give a significant boost to efficiency in metadata discovery coupled with classification accuracy in big industries typically associated with big and complex datasets. It indicates that the selection of any model for any specific task has a close relationship with the criticality AI has gained in optimization algorithms for metadata management systems.

Keywords: Metadata Management; Artificial Intelligence; Machine Learning; Data Discovery and Automation; Decision-Making; Analytical Work; Blockchain Technology; Customer Satisfaction.

Received on: 17/03/2024, Revised on: 10/05/2024, Accepted on: 01/07/2024, Published on: 01/09/2024

DOI: 10.69888/FTSCL.2024.000242

FMDB Transactions on Sustainable Computer Letters, 2024 Vol. 2 No. 3, Pages: 164-175

• Views : 140
• Downloads : 10

Download PDF