Showing papers on "Spatial analysis published in 1985"

Spatial Data Analysis by Example, Volume 1: Point Pattern and Quantitative Data

Abstract: Chapters are included on: analysis of the identifications of pattern; estimation of spatial intensity, mentioning the angle-count method; spatial auto-correlation; inter-type relations; and regression and auto-regression, including standard statistical material and examples of special cases applicable to spatial data analysis. Other CABI sites 

Measuring the unmeasurable

Abstract: Qualitative Spatial Data Analysis: A Compendium of Approaches.- A. Generalized Linear Models and Categorical Data.- Statistical Models for Qualitative Data.- Statistical and Scientific Aspects of Models for Qualitative Data.- Analysis of Qualitative Individual Data and of Latent Class Models with Generalized Linear Models.- Categorical Regression Models for Contextual Analysis A Comparison of Logit and Linear Probability Models.- Multivariate Contingency Table Analysis with NONMET: Basic Ideas.- Categorical Data Methods and Discrete Choice Modelling in Spatial Analysis: Some Directions for the 1980s.- B. Log-Linear Models.- Hybrid Log-Linear Models.- A Comparison of the Loglinear Interaction Model with Other Spatial Interaction Models.- Modelling Cross-Tabulated Regional Data.- C. Partial Least Squares and Lisrel models.- Systems Analysis by Partial Least Squares.- Recent Developments on Categorical Data Analysis by PLS.- Soft Modelling and Spatial Econometrics: Towards an Integrated Approach.- Structural Equation Models with Qualitative Observed Variables.- D. Multidimensional Qualitative Analysis.- Multidimensional Data Analysis for Categorical Variables.- Analyzing Activity Pattern Data Using Qualitative Multivariate Statistical Methods.- Unemployment and the Rise of National Socialism: Contradicting Results from Different Regional Aggregations.- Generalized Path Analysis for Mixed Geographical Data.- Order-Dependent Measures of Correspondence for Comparing Proximity Matrices and Related Structures.- A Survey of Qualitative Multiple Criteria Choice Models.- E. Fuzzy and Qualitative Structures.- A Linguistically-Based Regional Classification System.- Fuzzy Data Analysis in a Spatial Context.- Qualitative Structure Analysis of Complex Systems.- F. Discrete Choice Models and Dynamics Analysis.- Trends and Prospects for Qualitative Disaggregate Spatial Choice Models.- The Analysis of Panel Data for Discrete Choices.- Travel-Activity Behavior in Time and Space: Methods for Representation and Analysis.- Dynamic Analysis of Qualitative Variables: Applications to Organizational Demography.- Mathematical Specification of Transportation Models.- General Representational Formalisms and Search Procedures for Inferring Models from Categorical Data.- G. Synthesis.- Developing Trends in Qualitative Spatial Data Analysis.

Spatial autocorrelation of individual tree characteristics in loblolly pine stands

Abstract: Mathematical methods of assessing the levels of spatial autocorrelation in forest stands were identified. These methods were used to investigate spatial autocorrelation associated with tree characteristics such as product, defect, species class, and basal area. With the exception of the species classification, significant levels of spatial autocorrelation were not associated with any discrete variable. For individual tree basal area, the levels of spatial autocorrelation tended to be positive for low levels of competition, negative at intermediate levels of competition, and positive again at high levels of competition. Using measures of spatial autocorrelation, characteristics were assigned to individual trees in computer generated stands. These methods, applicable for discrete or continuous characteristics, assign the characteristics to individual trees depending on the spatial location of the individual tree and characteristics of its neighbors. FOREST SCl. 31:5 7 5-5 8 7. ADDITIONAL

Use of indicator variograms for an enhanced spatial analysis

Abstract: Flat variograms often are interpreted as representing a lack of spatial autocorrelation. Recent research in earthquake engineering shows that nearby field noise can substantially mask a prominent spatial autocorrelation and result in what appears to be a purely random spatial process. A careful selection of threshold in assigning an indicator function can yield an indicator variogram which reveals underlying spatial autocorrelation. Although this application involves use of seismic data, the results are relevant to geostatistical applications in general.

Spatial autocorrelation as a criterion for retaining factors in ordinations of geographic data

Abstract: Scientists often use factor analysis to reduce the number of dimensions needed to meaningfully describe the variation of species' abundances at a set of sample locations. Their assessments of appropriateness and adequacy of a factor solution for explaining observed geographic variation are limited to subjective considerations of numerical results and maps. Spatial autocorrelation analysis of the factor solution can be used to test the sufficiency of results objectively. This approach provides one way for investigators to determine how many factors should be retained and to evaluate the performance of those factors as descriptors of the spatial pattern.

Using Quadtrees to Represent Spatial Data

Abstract: Use of the quadtree data structure in representing spatial data is reviewed. The focus is on its properties that make it appropriate for applications in image processing. A number of operations in which the quadtree finds use are discussed.

Categorical Data and Choice Analysis in a Spatial Context

Abstract: This paper provides a comprehensive review of the techniques available for categorical data and discrete choice analysis in a spatial context. The first half of the paper reviews the class of models known as generalized linear models including log-linear models, linear logit regression models and latent class models. The second half of the paper focusses on techniques that may be used for categorical spatial choice analysis including utility-based models, multiattribute preference models and heuristic choice techniques.

COMPUTER-AIDED DATA ANALYSIS A Practical Guide

Abstract: Applications of Data Analysis. BASIC DATA ANALYSIS USING A COMPUTER. Initial Steps in Computer Data Analysis. Basic Statistical Analysis. Multivariate Data Analysis. Effective Use of Computerized Analysis Systems. ANALYSIS OF SPATIAL DATA WITH COMPUTER GRAPHICS. Fundamentals of Computer Plotting. Effective Use of Computer Plotting. Computer Plots as Aids to Data Analysis. Enhanced Display Techniques. INTRODUCTION TO ADVANCED ANALYSIS METHODS. Advanced Statistical Techniques. Procedures for Array-Oriented Data. Physical Models and Data Interpretation. Appendixes. References. Glossary.

Geographic and Spatial Data Management and Modeling

Abstract: A geographic information system is a computer based technology for storing and using spatial data. Many alternative methodologies exist for managing the three general types of spatial data: areal data, terrain data, and network data. The primary uses for spatial data, mapping and modeling, have been applied extensively in the area of water resources.

Strategies for Associating Data and Location in a Geographic Information System.

Abstract: : Much of the existing work in the area of Geographic Information Systems (GIS) treats spatial objects, e.g. points, lines, and regions, as the primary entities of interest. In that approach, descriptive information is associated directly with each of these data items. This paper explores the feasibility of implementing an alternative design which uses Location Data Sets and Location Predicates as the basic entities managed by a Location Data Management System (LDMS). A major advantage of the proposed approach is its suitability for automatic enforcement of data consistency across multi-scale geographic entities. The central idea of the Location Data Set approach is that spatial data should be directly associated with locations rather than named regions or points. The relationships between geographic entities and data values may then be derived through the intermediate relationship of shared location. It is envisioned that each type of data which is distributive in nature would be stored in a separate set. Data values associated with conventional points, lines, and regions would then be merely restrictions on these global data sets. This is similar to the way in which the external views of a database represent a subsetting of the global data. The paper includes a survey of fifteen selected GIS implementations and existing work relevant to identified implementation obstacles. (Author)

A prospective case for a national land data system: ten years later

Abstract: Geographic information systems are emerging as powerful tools for the handling of spatial data. In the past, a geographic information system was often designed to meet the needs of a specific problem, and the data capture, management, and analysis functions were often restricted to the unique characteristic of specific data sets. More recently, systems have been designed for generic data types and functions to provide much greater flexibility and a wider range of applications. The emergence of these systems, along with a new national data base being developed jointly by the Bureau of the Census and U.S. Geological Survey offer the opportunity to establish a solid foundation for a National Land Data System.

Spatial trend analysis and uncertainty estimates of acid precipitation data in Ontario

Abstract: The use of the Simple Kriging and Universal Kriging methods to estimate the spatial trend in the spatial data is examined. By following the approach of the Simple Kriging methods, a new Modified Kriging method has been developed for better presentation of the spatial pattern of spatial data and error estimation. Results of comparison indicate that the Modified Simple Kriging method can provide better interpolation and error estimation for the mathematical functions. Although the results of practical interpolation obtained by using the three methods are quite comparable, the error estimates by the Simple Kriging and Universal Kriging methods are greater than those of the modified Simple Kriging method and the one standard deviation confidence belts are overlapping each other. This may indicate that the error estimations from the Simple Kriging methods are too great.The Universal Kriging method provides the greatest error estimation among these three methods. Although those results indicate that there is no advantages of using the Universal Kriging method, further study on the proper formula for the generalized covariance function should be carried out. The heart of the Universal Kriging method is to find the correct covariance function, since the isopleths and the error of estimation are verymore » sensitive to it.« less

Technical support for creating an artificial intelligence system for feature extraction and experimental design

Abstract: Techniques for classifying objects into groups or clases go under many different names including, most commonly, cluster analysis. Mathematically, the general problem is to find a best mapping of objects into an index set consisting of class identifiers. When an a priori grouping of objects exists, the process of deriving the classification rules from samples of classified objects is known as discrimination. When such rules are applied to objects of unknown class, the process is denoted classification. The specific problem addressed involves the group classification of a set of objects that are each associated with a series of measurements (ratio, interval, ordinal, or nominal levels of measurement). Each measurement produces one variable in a multidimensional variable space. Cluster analysis techniques are reviewed and methods for incuding geographic location, distance measures, and spatial pattern (distribution) as parameters in clustering are examined. For the case of patterning, measures of spatial autocorrelation are discussed in terms of the kind of data (nominal, ordinal, or interval scaled) to which they may be applied.

Taking into account spatial dependence in multivariate analysis

Abstract: R is a valuable tool to develop statistical methods using its package structure and the open sources. Indeed, it allows combining classes and methods from different packages in order to generalize their own capabilities. The present topic deals with the relationship between classes and methods from both packages.spdep and ade4, which implement methods for spatial data analysis (mainly lattice/area style) and multivariate analysis respectively. Combining objects from both packages leads to the development of a new statistical method to take into account spatial dependence in multivariate analysis. spdep is a package defined by Bivand as “a collection of functions to create spatial weights matrix W from polygon continuities, from point patterns by distance and tessellations for permitting their use in spatial data analysis”. Two classes of objects have been established: “nb” a list of vectors of neighbour indices, and “listw” a list containing a “nb” member and a corresponding list of weights for a chosen weighting scheme. The package contains in addition many functions to test spatial autocorrelation and estimate spatial autoregressive model but it only supports univariate data analysis. ade4 is precisely a package dedicated to multivariate analysis of Ecological and Environmental Data in the framework of Euclidean Exploratory methods. It is a collection of functions to analyse the statistical triplet (X,Q,D) where X is a data set of variables or a distance matrix; Q and D are diagonal matrices containing column and row weights, respectively. The singular value decomposition of this triplet gives principal axes, principal components and, column and row coordinates. All these elements are gathered in a list defining the duality diagram class called “dudi”. Each basic statistical method corresponds to the analysis of a particular triplet. For instance, ade4 implement the principal component analysis on correlation matrix via a function called “dudi.pca”. In that case, X is a table