Vision No. 1 July 2006
Welcome to Vision, the newsletter of VSNi.
Well, it has been a while since the last newsletter and a quick glance at this one will tell you why. Http://monstersessay.com/custom-essay-service/ New corporate branding, new website, GenStat 9th edition and ASReml 2 have all been keeping us a little busy.
GenStat 9th edition is From eagerly awaited, containing many improvements that our fantastically loyal user groups have suggested. From new interface and help functions to more statistics and graphics menus – each development has been included for one reason – you.
We’re also delighted to be launching ASReml 2, and yes it has taken a while. ASReml is rightly known as one of the best Reml packages around, users rely on it, so we needed to get ASReml 2 right. And we have. Take a look at the new releases section.
Talking of ASReml 2, Dr Arthur Gilmour, one of the Reml gurus, will be running a workshop for us in August. This is a rare chance to develop your Reml skills alongside the master, why not join us? Many places are already taken, so book now to avoid disappointment.
Finally, congratulations to Professor Nelder on his IBS Honorary Life Member Award.
Oh – and don’t forget to visit our new website at: www.vsni.co.uk
GenStat 9th Edition
We are delighted to announce the release of GenStat 9th Edition.
“The new statistical features and interface enhancements in the 9th Edition maintain VSNi’s reputation for providing state-of-the-art statistics in a very convenient framework” says Prof. Roger Payne, Chief Technical Officer of VSN International.
New statistical features include complex surveys, diversity statistics, split-line regression and Bayesian analyses using the DEMC algorithm. Permutation and exact tests can be done for regression, generalized linear models, analysis of variance and nonparametric analyses. Other anova enhancements include screening tests for unbalanced data and multitiered analyses.
“The release further enhances GenStat’s existing strengths in anova and generalized linear models, while adding new facilities for areas such as ecology, microarray and survey analysis.”
New, more efficient algorithms are provided for analysing hierarchical generalized linear models, and examples are included from the new CRC book by Youngjo Lee, John Nelder and Yudi Pawitan. There are also many enhancements to the general user interface, including the ability to include Greek letters and mathematical symbols in output, multi-tabbed spreadsheets and new tree-view facilities to access data and windows.
Key new features in GenStat for Windows™ – 9th Edition
GenStat Server has been upgraded to Release 9.1
- Includes 2 new directives, 64 new procedures and 2 new functions
Improved interface and help facilities
- New docked data pane and window navigator
- Improved searchable lists of example programs and procedure source
- Run an R script from GenStat
New supported file formats
- Support for SAS catalogues defining value labels
- Support for JMP 4 and 5, Paradox 7-9, SPSS 13-14 file formats and R gzip compressed data frames
- Import files from ZIP archives
New statistical and graphics menus
- Species abundance plot menu for producing rank/abundance, ABC or k-dominance plots
- Lorenz curve menu for producing a Lorenz curve and calculating the gini coefficient
- Transformed axes
- Diversity indices menu for calculating measures of diversity and jackknife confidence intervals
- T-test menu includes facility for calculating permutation tests
- Nonparametric correlations menu for calculating Spearman’s rank correlation and Kendall’s rank correlation coefficient
- Cochran’s Q Test menu
- Species abundance models menu for fitting a range of distributions and models to describe species abundance data
- Split-line regression menu for fitting a regression model consisting of two straight line segments (a split-line or broken-stick model) to the data
- Linear and generalized linear regression menus include facility for calculating permutation tests
- Create survey weights menu to generate weights to be used in survey analysis
- Modify survey weights menu to modify weights for observations in survey analysis
- Calibration weighting to perform calibration estimation of survey data
- General survey analysis menu to calculate estimates from surveys together with the asymptotic standard errors
Analysis of Variance
- Menus include facility for calculating permutation tests
- Parallel ANOVA menu run a large number of Analysis of Variances in parallel for a set of units that have multiple measurements on each unit
- Prediction menu for Hierarchical Generalized Linear Models
New spreadsheet facilities
- Spreadsheets can be stored in multi-paged books where each spreadsheet is represented by a different page within a book
- Data can be appended into a spreadsheet simultaneously from multiple files
- Data can be appended from multiple worksheets from an Excel file to a GenStat spreadsheet
- Menu to randomize rows within a spreadsheet
- Drag and drop columns between sheets
- Improved replace facilities to allow multiple replacements of text within a single cell
ASReml version 2
ASReml is a statistical package that fits linear mixed models using Residual Maximum Likelihood (REML). Linear mixed effects models provide a rich and flexible tool for the analysis of many data sets commonly arising in the agricultural, biological, medical and environmental sciences. Typical applications include the analysis of (un)balanced longitudinal data, repeated measures analysis, the analysis of (un)balanced designed experiments, the analysis of multi-environment trials, the analysis of both univariate and multivariate animal breeding and genetics data and the analysis of regular or irregular spatial data.
ASReml is a joint project of NSW Department of Primary Industries and Rothamsted Research. It provides a stable platform for delivering well established procedures while also delivering current research in the application of linear mixed models. The strength of ASReml is the use ncaa of the Average Information (AI) algorithm and sparse matrix methods for fitting the linear mixed model. As a result, ASReml analyses large and complex data sets more efficiently than other packages.
Uniquely efficient and fast algorithms for mixed model analysis, saving you considerable time and effort:
- Faster equation ordering algorithm
Handles large data sets (of 100,000 or more observations/effects):
- More transformations
- Data read from multiple files
Handles complex linear mixed models:
- Allows for direct fitting of cubic smoothing splines with user-specified knot points
- Facilitates multi-environment trials for the analysis of plant or crop improvement data
- Analyses univariate and multivariate breeding and genetics data
- Encourages innovative modelling of longitudinal data
- Generalized Linear Models
Handles complex variance structures based on direct products as well as traditional variance component models:
- Supports a wide range of variance models for spatial analysis
- Matern isotropic covariance functions
- Extended Factor Analytic in combination with other variance models
- New pedigree options
Advanced Analysis of Variance procedures suitable for unbalanced mixed models
- Wald statistics / Kenward & Roger adjustments
- Slash (/) operator in model specification (nested models)
Advanced procedures for fitting difficult models
- Improved ability to estimate variance matrices when the matrix is effectively singular by using EM steps
User Interface: Improvements to user interface, graphics, and online documentation
- Templates: automatic generation of program templates from data files
- Extended facilities for paths and loops in the program file
- Additions to Prediction, including graphical output
As you will have probably gathered by now, we have launched our new website. Click here
to take a look. Actually, some of you have already visited whilst we were trialling and testing – and thank you for your positive feedback.We are really pleased with the development. It looks cleaner, information is much easier to find, as are the downloads and relevant product support pages.As with our product development, we are more than happy to hear your views so please email us
to let us know what you think.
VSN International bring you a unique opportunity to further your knowledge and experience of ASReml with the help one of the ASReml Personalized creators, Dr Arthur Gilmour, Principal Research Scientist of New South Wales Department of Agriculture, Australia. During his forthcoming visit to the UK we are pleased to announce that Professor Gilmour will be running an intermediate ASReml workshop in Hemel Hempstead on 29th and 30th August 2006.
Who Should Attend?
Anyone who is already familiar with the ASReml system and wishes to increase their knowledge and use of ASReml analyses.
£790 per commercial participant (ex VAT) £590 per academic participant (ex VAT)
Aim of Workshop
To give participants insight into the concepts underlying ASReml and hands-on experience with ASReml relevant to their own applications.
Session Topics and Learning Objectives
- ASReml 2 – How to recognise and use new facilities.
- Animal models – To understand the concepts of variance modelling.
- ANOVA – How to use conditional F statistics.
- Exploring your data – How to use descriptive statistics in ASReml.
- Multivariate models – To understand and use the options for singular variance structures.
- Multienvironment trials – How to use Factor Analytic models.
- Prediction – To understand the concepts behind prediction and common reasons for failure.
- Analysing your data – How to select appropriate variance models.
Places are already booking fast, so contact us today at training to book your place.
Congratulations to Professor John Nelder
VSNi would like to congratulate Professor John Nelder on his International Biometric Society Honorary Life Membership, to be awarded at this year’s conference held in Montreal, Canada. This prestigious award can only complement his already extensive collection of honours, including the Royal Statistical Society’s Guy Medal in Gold awarded last year. John’s contributions include the simplex minimization algorithm, the theory of generally-balanced experimental designs, the methodology of generalized linear models and, more recently, their generalization to hierarchical GLMs. He is also the originator of the two widely used and highly respected statistical packages, GLIM and of course, our very own GenStat.
Warmest congratulations John.