Grains and field crops, Biosecurity, pests, weeds and diseases
Department of Agriculture
Notes for a workshop on pattern analysis of germplasm data ; to be held at the Pathumk Thani Rice Research Centre, May-June 1989. At head of title : Australian Co-operation with the National Agricultural Research Project Thailand.
Pattern analysis is a term which covers a very large area of statistical analysis for multivariate data sets. A full understanding of pattern analysis would require years of study of mathematical concepts and computing algorithms. Fortunately, pattern analysis methods can now be used without all this specialist knowledge. This is possible because of the availability of computer programs such as PATN, which is a powerful interactive program running o" micro computers. However,such powerful programs are easily misused. If researchers are to make good use of such programs, they should have some understanding of the concepts of pattern analysis, and the options available.
The purpose of these notes is to provide this basic descriptive understanding of the mathematical routines used in pattern analysis. We hope that this will help people to decide when pattern analysis might be useful, and how to interpret the results. Mathematical formulas are avoided as far as possible. However, pattern analysis has a great many 'special words' with specific meanings. Use of these terms is unavoidable. In a way, pattern analysis begins where normal statistical methods end. Some knowledge of basic statistics is assumed. These notes are far from complete.
They are to be used in conjunction with lectures, and hands-on demonstrations of the PATN program on users's data sets. For a fuller coverage of the subject, some of the references should be consulted.
Number of Pages
Wallace, J F, and Australian Co-operation with the National Agricultural Research Project, Thailand.. (1989), Pattern analysis : a descriptive introduction. Department of Agriculture, Perth. Report.
This file is 11.4 MB. Files over 3MB may be slow to open. For best results, right-click and select "save as..."