Privacy-Preserving Decision-Making for Digital Transformation with Divide-and-Conquer Data Aggregation

In the Digital Transformation (DX) era, anonymization is important to protect data privacy, but it often faces limitations in maintaining data quality and complying with privacy regulations within organizations. This paper introduces a divide-and-conquer retention and aggregation approach for data protection in decision-making. The approach involves processing data from the information owner and making decisions based on a machine learning scheme. A divide-and-conquer approach to dividing important attributes for aggregation was proposed. The proposed approach was applied to loan application data. The test results showed aggregated decision-making has a higher correlation and accuracy than non-aggregated decision-making. The result showed that the aggregated decision value is more similar to the original value even after masking the original value with the aggregated value.