ITR/SMALL/Scientific Frontiers: Task-Specific Data Reduction and Mining in Spatial-Temporal Domains

ITR/小/科学前沿：时空域中特定任务的数据缩减和挖掘

• 批准号: 0219736
• 负责人: Zoran Obradovic
• 金额: $ 21.01万
• 依托单位: Temple University
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2002
• 资助国家: 美国
• 起止时间: 2002-08-15 至 2006-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0219736&HistoricalAwards=false
• 关键词: ITR; SMALL; Scientific; Frontiers; Task

Advances in earth sciences have spurred a large growth in spatial-temporal databases with multidimensional attributes. Unfortunately, theimprovement in data collection techniques has not been accompaniedwith the efficient methods that analyze large databases, which currentlylimits their use. To address the need for facilitating large-scaleknowledge discovery at a wide range of spatial-temporal datasets theobjective of the proposed project is development of novel techniquesfor (1) task-specific data reduction in spatial-temporal databases; and(2) efficient knowledge discovery on reduced spatial-temporal data.The first task builds on the observation that the amount of collected datais not necessarily correlated with specific knowledge discovery tasks,and that real-life datasets are often highly redundant. The second taskaddresses a number of open questions for more efficient mining ofreduced spatial-temporal datasets including exploring trade-offsbetween model complexity and data reduction for efficient learning,exploiting residual correlations for prediction, and task-specificpartitioning of nonstationary spatial-temporal data. The proposedtechniques will be evaluated on multi-source datasets with variouscomplexities. This project will result in guidelines for effectivedissemination of spatial-temporal data sets to wider research audienceand algorithmic tools for more successful knowledge discovery from suchdata.

地球科学的进步促进了具有多维属性的时空数据库的大量增长。不幸的是，数据收集技术的改进并没有伴随着分析大型数据库的有效方法，这限制了它们的使用。为了满足在广泛的时空数据集上促进大规模知识发现的需求，拟议项目的目标是开发新的技术，用于(1)时空数据库中特定任务的数据缩减；(2)基于约简时空数据的高效知识发现。第一个任务建立在这样的观察之上，即收集的数据量不一定与特定的知识发现任务相关，而且现实生活中的数据集通常是高度冗余的。第二个任务解决了一些开放的问题，以更有效地挖掘减少的时空数据集，包括探索模型复杂性和数据减少之间的权衡，以实现高效学习，利用残差相关性进行预测，以及非平稳时空数据的任务特定划分。所提出的技术将在具有不同复杂性的多源数据集上进行评估。该项目将产生有效传播时空数据集的指导方针，以更广泛的研究受众和算法工具，以便更成功地从这些数据中发现知识。