Supporting Multiple Time Granularities in Query Evaluation and Data Mining

点击翻译按钮获取中文摘要

The objective of this project is to provide a flexible framework for supporting multiple time granularities. The targeted use of the framework is for automatic evaluation of user queries and for discovering temporal patterns (i.e., data mining) in an environment where either the user queries or temporal patterns involve granularities that do not match the granularity of the stored data. The basic idea is to add the necessary functionalities so that the database system is able to understand and reason about information involving multiple time granularities. Algorithms for efficiently evaluating user queries and discovering temporal patterns will also be investigated, and an experimental prototype will be built. The impact of this research should be twofold: Firstly, the proposed framework should provide the basic tools for database systems to support flexible user queries, which would free users from the unnecessary tasks of writing complex queries or changing their pattern descriptions to fit the specific granularity used to store temporal data in the database. Secondly, the optimization techniques explored should provide strategies and algorithms for efficient evaluations of queries and discovery of temporal patterns. The proposed experimental work should provide insights into the subject and into the practicality of the proposed theory as well as methods.