Privacy-Preserving and Action-Event Sequence Data Mining and Advanced Data Structures for Efficient Heuristic Search

机器学习和数据挖掘技术的最新进展为将这种技术应用于政府和机构数据库提供了机会。目前的一个挑战是了解个人的一般行为，以便政府和机构能够针对特定个人采取适当行动，同时保护所有人的隐私。这项研究项目的一个目的是解决作为有关个人或实体的事件序列收集的数据所面临的这一挑战。对于项目1，我将设计和评估用于保护隐私的数据挖掘的匿名化方法。*其次，我将设计和评估基于各种匿名序列的已知效用来为特定事件序列选择高效用操作的技术。对于项目2，我将继续研究为高效的启发式搜索设计新的数据结构。

Recent advances in machine learning and data mining techniques present the opportunity of applying such techniques to governmental and institutional databases. One current challenge is to learn about the general behavior of individuals so that appropriate actions can be taken for specific individuals by governments and institutions while protecting the privacy of all. One purpose of this research project is to address this challenge for data collected as sequences of events concerning individuals or entities. For Project 1, I will devise and evaluate methods for anonymization for privacy preserving data mining.******Secondly, I will devise and evaluate techniques to find high-utility actions to choose for particular event sequences based on the known utilities of diverse anonymized sequences. For Project 2, I will continue research on a promising thrust towards devising new data structures for efficient heuristic search.