Graphical Analysis of Multi-Response Data

Preface, Acknowledgements, Part A, 1 Introduction to Part A, 2 The example, 3 Styles of analysis, 3.1 Preliminary considerations, 3.2 Scatter Plot Matrices, 3.3 Within environment analyses for single attributes, 3.4 Overlap plots, 3.5 Profiles across environments and B-confidence, 3.6 Grouping the data, 3.7 Smoothing the profiles, 3.8 Profiles across attributes, 3.9 Choice of expression, 3.10 This account in a larger framework, Part B, 4 Introduction to Part B, 5 Overall behaviour: SPLOMs across attributes and across environments, 6 Semigraphical comparison of genotypes for each attribute-environment pair, 7 Profiles of individual genotypes and groups, 8 SPLOMs for genotype groups, 9 Re-attribution of the responses on the 43 genotypes, 9.1 Our specific example, 9.2 Styles of re-attribution, 10 Semigraphical displays for the re-attributed data, 11 Re-environmenting, 11.1 Two remarks, 11.2 Overview of position, 12 Re-expression, Part C, 13 Introduction to Part C, 14 Global aspects of the data, 14.1 Ordering of genotypes, 14.2 Linear ordering of genotypes, 14.3 Dispersed inheritance, etc, 14.4 Characterizing and ordering of environments, 14.5 Ordering of attributes, 15 Data laundry, 15.1 Seeking exotics, 15.2 Winsorization of anomalies, 16 Choices of expression, 16.1 Comparative scaling, 16.2 Comparative responses, 16.3 Re-expression of individual attributes, 17 Seeking exotic values, 17.1 SPLOMs, 17.2 Alternative displays, 17.3 Correlations among environments, 18 Local analyses and displays, 18.1 Local analysis within environments, 18.2 Adjustment of structured sets of s2 values, 18.3 Choice of critical levels, 18.4 Benjamini-Hochberg (FDR) procedure, 18.5 Overlap plots, 18.6 Dealing with ties in the BSD procedure, 18.7 Profiles across environments and B-confidence, 18.8 Some plausible developments for the BSD, 19 Combined analyses, 19.1 Analysis over environments, 19.2 Analyses over attributes, 20 Grouping and labelling genotypes, 20.1 Grouping, 20.2 Labelling, 21 Idiolinkage, nearest and centroid, 21.1 Nearest, 21.2 Nearest algorithm, 21.3 Centroid, 21.4 Mean squared distance, 22 Other blended forms; xpanded idiolinkage, 22.1 Other blended forms, 22.2 Xpanded linkage, 22.3 Generalized xpansion, 22.4 Techniques so far proposed, 23 Robust forms of idiolinkage, 23.1 Flexible Winsorization, 24 Idiomax, 24.1 Expansion, 24.2 Plausible scope of exploration, 25 Robust sphering, 25.1 Robust orthogonalization, 25.2 Simplification, 25.3 Robust scaling, 25.4 Comments, 25.5 Relative importance?, 26 A side issue and more careful sphering, 26.1 Pattern of analysis, 26.2 Choice of scaling, 26.3 A possible exception, 26.4 Postcomment, 27 A suggestion about ‘G in or out’, 27.1 Another suggestion, 28 Detrivialization of one- and two-way tables: an, introduction, 28.1 The one-way case, 28.2 The two-way case, 29 Later attributes and subtables, 29.1 The second attribute, 29.2 The third attribute, 29.3 Later attributes, 29.4 Combination, 29.5 Postcomment, 30 Sphering the soybean data, 30.1 Ordering of attributes, 30.2 Robust orthogonalization, 30.3 Robust scaling, 31 Grouping the soybean genotypes, 31.1 First-stage grouping, 31.2 Second-stage grouping, 31.3 Group responses, 31.4 Alternative analyses, 31.5 Partitioning variability within groups, 32 Ordering the environments, 32.1 Greedy close ordering of environments, 33 One approach to plotting, 33.1 Smoothing in general, 33.2 Group 4 and attribute 1 as an example, 33.3 Looking at the seven groups for attribute 1, 33.4 Looking at the seven groups for all attributes, 34 Smoothing short sequences (of perhaps 6 to 18 values), 34.1 3R 3pR E2, 34.2 3+R E2, 35 Presentations across attributes, 36 Choice of expression: generalities, 36.1 Matching re-expressions, 36.2 Purposes of choice of expression, 36.3 Rarity of conflict between these goals, 36.4 Tools for diagnosis, 36.5 Combining Q-diagnoses, 36.6 Breaking down the example, 36.7 Weighted regression, 37 Removable inhomogeneity of variability, 37.1 Replication within environments (e.g. 2x40), 37.2 Replication between environments (e.g. 8x40), 37.3 Inhomogeneity of interaction size (e.g. 8x40 or 8x18), 37.4 Variability between replications within environments, 37.5 The special case of two versions, 37.6 Interaction variability, 38 Removable non-additivity, 38.1 Extended fits, 38.2 Choice of expression to promote additivity of response, 38.3 General case of several versions, 39 Linearity of response, 40 Relative importance and combination, 40.1 Bowstring plots, 41 Interpretation and hybridization, Appendices, Appendix 1, Appendix 2, Appendix 3, Appendix 4, Appendix 5, Appendix 6, Appendix 7, Appendix 8, Appendix 9, Appendix 10, Appendix 11, Appendix 12, Appendix 13, Appendix 14, Appendix 15, Appendix 16, Appendix 17, Appendix 18, Appendix 19, Appendix 20, Appendix 21, Appendix 22, References, Author Index, Subject Index