Multi-Model Machine Learning for Identifying Abnormal Data Activities and Securing Cloud Data

In today’s data-driven digital world, data classification and data loss prevention are the foundational techniques for safeguarding sensitive data and mitigating data breaches. Data classification organizes data into appropriate categories, such as confidential, private, sensitive, and public, while Data Loss Prevention (DLP) focuses on preventing unauthorized disclosure of sensitive information. The paper utilizes machine learning (ML) to enhance both data classification and DLP for cloud data protection. It proposes a multi-model ML pipeline in which the first model classifies the sensitivity levels of data, while the second model uses this output, along with original dataset features, to predict the threats to sensitive data and enhances DLP by identifying unusual cloud activity.