The goal of the proposed project is to design efficient algorithms applicable to security and computational issues that are core concerns of statistical agencies and database vendors. Results obtained from the project will be implemented to test theoretical models as well as to advance technology transfers. Building on the PI's completed work on statistical data security, a number of specific research directions are pursued to further determine the computational complexity of detecting and protecting sensitive information in highly useful types of statistical data bases. These directions involve techniques from a variety of algorithms areas including combinatorial optimization, dynamic algorithms, graph algorithms, linear algebraic algorithms, and online competitive analysis. It is expected that algorithmic insights and techniques obtained from this project will be useful in various application areas in addition to statistical data security, such as relational databases and secret sharing schemes.
