Techniques for dealing with missing values in classification

被引：0

作者：

Liu, WZ

White, AP

Thompson, SG

Bramer, MA

机构：

[1] Univ Portsmouth, Artificial Intelligence Res Grp, Dept Informat Sci, Milton PO4 8JF, Hampshire, England

[2] Univ Birmingham, Sch Math & Stat, Birmingham B15 2TT, W Midlands, England

来源：

ADVANCES IN INTELLIGENT DATA ANALYSIS: REASONING ABOUT DATA | 1997年 / 1280卷

关键词：

missing values; dynamic path generation; intelligent data analysis; inductive learning; knowledge discovery; data mining; machine learning;

D O I：

暂无

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

A brief overview of the history of the development of decision tree induction algorithms is followed by a review of techniques for dealing with missing attribute values in the operation of these methods. The technique of dynamic path generation is described in the context of tree-based classification methods. The waste of data which can result from casewise deletion of missing values in statistical algorithms is discussed and alternatives proposed.

引用

页码：527 / 536

页数：10

共 18 条

[1]

Breiman L., 1984, Classification and Regression Trees, DOI DOI 10.2307/2530946

[2]

Clark L.A., 1992, STAT MODELS S, P377

[3]

FRIEDMAN HF, 1996, P 13 NAT C ART INT, P717

[4]

Hunt E. B., 1966, EXPT INDUCTION

[5]

Hunt E. B., 1962, Concept learning, an information processing problem

[6]

Kass G. V., 1980, Applied Statistics, V29, P119, DOI [DOI 10.2307/2986296, 10.2307/2986296]

[7]

KONONENKO I, 1984, EXPT AUTOMATIC LEARN

[8]

Liu W. Z., 1994, Machine Learning: ECML-94. European Conference on Machine Learning. Proceedings, P391

[9]

LIU WZ, 1994, MACH LEARN, V15, P25, DOI 10.1023/A:1022609119415

[10] Machine learning techniques in early screening for gastric and oesophageal cancer [J].

Liu, WZ ;

White, AP ;

Hallissey, MT ;

Fielding, JWL .

ARTIFICIAL INTELLIGENCE IN MEDICINE, 1996, 8 (04) :327-341

← 1 2 →