POWRE: Effects of Systematic Data Error on Inductive Machine Learning Algorithms

POWRE：系统数据错误对归纳机器学习算法的影响

• 批准号: 9806218
• 负责人: Andrea Danyluk
• 金额: $ 7.44万
• 依托单位: Williams College
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1998
• 资助国家: 美国
• 起止时间: 1998-09-01 至 2000-02-29
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9806218&HistoricalAwards=false
• 关键词: POWRE; Effects; Systematic; Data; Error

EIA-9806218Danyluk, AndreaWilliams CollegePOWRE/CISE: Effects of Systematic Data Error on Inductive Machine Learning AlgorithmsProposed research is a change of direction for the PI, and provides her with an opportunity to switch from applied research to theoretical aspects of machine learning. The objective is to study the effects of systematic data error on inductive machine learning algorithms. The work proposed will begin to characterize the effects of systematic error both analytically and empirically. It will incorporate development of a model for analyzing worst case effects of data error on decision tree learners. One of the objectives of the study is to increase the involvement of undergraduates in research and provide them with an opportunity to obtain active research experience.

EIA-9806218Danyluk， andrewilliams college power /CISE：系统数据误差对归纳机器学习算法的影响拟议的研究是PI方向的改变，并为她提供了从应用研究转向机器学习理论方面的机会。目的是研究系统数据误差对归纳机器学习算法的影响。提出的工作将开始从分析和经验两方面描述系统误差的影响。它将包含一个模型的开发，用于分析数据错误对决策树学习器的最坏情况影响。这项研究的目的之一是增加本科生对研究的参与，并为他们提供获得积极研究经验的机会。