04:30 PM in LAEB Room 1268
K. C. S. Pillai Memorial Lecture
B.L.S. Prakasa Rao, Indian Statistical Institute
Spatial Data Analysis
Abstract
In order to study the relationship between one variable and
another variable, classical statistical methods range from scatter plot
techniques for visual exploration of data to regression analysis and
more sophisticated techniques. Regression techniques might include
some explanatory variables or covariates other than the two variables
under consideration. However one factor which may not be incorporated
in the standard regression analysis is the spatial configuration
of the observational units. For instance, in the study of relationships
between voting pattern or support for a political party and the level
of income, the party support and per capita income in one region
may be related to those in the neighboring constituencies in addition
to the relationship with each other within the region.
In this talk which is of introductory nature, we will briefly describe
four types of spatial data: point patterns, spatially continuous
data, aerial data and spatial interaction data. Some methods of
statistical analysis for such data will be presented. Application of
methods of nonparametric inference and statistical analysis of random
fields to such data will be discussed.
References:
1) Statistical analysis of spatial point patterns, Diggle (1983)
2) Effective observation of random fields, Nather (1985)
3) Spatial statistics and imaging (Ed., Possolo)
4) Statistics for spatial data, Cressie (1993)
5) Interactive spatial data analysis, Barley and Gatrell (1995)
References:
1) Statistical analysis of spatial point patterns, Diggle (1983)
2) Effective observation of random fields, Nather (1985)
3) Spatial statistics and imaging (Ed., Possolo)
4) Statistics for spatial data, Cressie (1993)
5) Interactive spatial data analysis, Barley and Gatrell (1995)
