STOCHASTIC DISTURBANCE SPECIFICATION AND ITS IMPLICATIONS FOR LOG-LINEAR REGRESSION

被引：27

作者：

HAWORTH, JM ^{[1
]}

VINCENT, PJ ^{[1
]}

机构：

[1] UNIV LANCASTER,DEPT GEOG,LANCASTER LA1 4YW,ENGLAND

来源：

ENVIRONMENT AND PLANNING A | 1979年 / 11卷 / 07期

关键词：

D O I：

10.1068/a110781

中图分类号：

X [环境科学、安全科学];

学科分类号：

08 ; 0830 ;

摘要：

Model specification is a crucial factor in regression. In log-linear models, although the estimated relationship represents the conditional expectation of ln Y, the antilogarithm does not give the conditional expectation of Y. This has implications in spatial analysis. -Authors

引用

页码：781 / 790

页数：10

共 18 条

[1]

Aitchison J, 1957, LOGNORMAL DISTRIBUTI

[2]

BAXTER RS, 1976, COMPUTER STATISTICAL

[3]

Cliff A.D., 1973, SPATIAL AUTOCORRELAT

[4]

DRAPER N, 1966, APPLIED REGRESSION A

[5]

EBDON D, 1977, STATISTICS GEOGRAPHY

[6]

FERGUSON RI, 1977, CONCEPTS TECHNIQUES, V15

[7]

Goldberger AS, 1968, ECONOMETRICA, V35, P464

[8]

Goldfeld SM, 1972, NONLINEAR METHODS EC

[9] ACQUIRING SPATIAL INFORMATION [J].

GOULD, P .

ECONOMIC GEOGRAPHY, 1975, 51 (02) :87-99

[10] IS STATISTIX-INFERENS GEOGRAPHICAL NAME FOR A WILD GOOSE [J].

GOULD, P .

ECONOMIC GEOGRAPHY, 1970, 46 (02) :439-448

← 1 2 →