Secure and Efficient k-NN Queries

Area of Science:

• Computer Science
• Financial Technology
• Cryptography

Background:

• Ranking and best match queries are crucial for navigating large datasets.
• k-Nearest Neighbors (k-NN) queries identify the k closest data points to a query, with broad applications.
• The financial sector requires secure methods for querying sensitive investment portfolio data.

Purpose of the Study:

• To develop a secure k-NN computation protocol for distributed multi-party environments.
• To address the challenge of querying sensitive financial data while preserving privacy.
• To incorporate domain semantics into secure k-NN computations for financial applications.

Main Methods:

• Development of a novel secure k-NN computation protocol.
• Implementation of the protocol in a distributed multi-party setting.
• Evaluation of the protocol's efficiency and security in financial portfolio matching scenarios.

Main Results:

• The proposed secure k-NN protocol demonstrated high efficiency in experimental tests.
• The protocol successfully enabled secure k-NN queries on sensitive financial data.
• Domain semantics were effectively integrated into the secure computation framework.

Conclusions:

• The developed protocol offers an efficient and secure solution for k-NN queries in the financial sector.
• This work advances privacy-preserving data analysis in distributed financial environments.
• The protocol's efficiency and security make it suitable for real-world applications involving sensitive investment data.