Faculty of Agricultural Sciences

Today I received the formal letter from Århus University concerning the fusion of our research institute with the university per January the first. They wrote, that I will receive another letter in the near future from the Ministry of Food, Agriculture and Fisheries concerning the formal details about my reassignment. There are still almost 2 weeks until I receive my salary ;-) from the University.

In Danish the name of the new faculty is Det Jordbrugsvidenskabelige Fakultet, which means that we are able to maintain the three letter acronym DJF. I don't know the new acronym in English, it seems that the old one DIAS will have to be changed.

Based on new name there seems to be some interesting possibilities:-)

I still the recall when we moved from Copenhagen to Foulum. An officiel postcard was designed, which was meant to make it easy to send information to our international collaborators. The postcard showed a map of Denmark with a marker for Copenhagen and Foulum, and the title:

We have mowed

Since then I always imagine an army of researchers walking across Denmark, each with his own lawn mower...

Another nasty question

I think the single most frequent question during statistical consultancy here at Foulum is:

How do I calculate the standard eror on the original scale when I have made an analysis on transformed values.

Back in 1998 I wrote a short note adressing this problem together with Asger Roer Pedersen (now an employee at the National Envirenmental Institute).

Unfortunately, people have trouble finding it, so just to refresh

Jørgensen, E. & Pedersen A.R. (1998 ). How to obtain those nasty standard errors from transformed data - and why they should not  be used. Biometry Research Unit - Internal report 7. Danish Institute of Agricultural Sciences. pp 20

Or, via my publication page: Erik Jørgensen's Publications

A new blog is taking form

I have gathered a lot of examples about the use of Bayesian Networks and Markov Processes during the last 15 years. That is why I have startet a new blog : Bayesian Networks and Markov Processes in Agriculture. The examples have been used for different lectures, presentations and workshops. At the moment, the examples are more or less scattered on different obsolete servers and some of them on my own hard disk.

My plan is to transfer the examples to the blog, when I have some spare time. In addition examples of current research will be included

Until now I have only made a few postings, mainly based on examples in Bedre Beslutninger med Bayesianske Netværk and Potential Application Areas for Bayesian Networks within Animal Production.

In addition to Bayesian Networks, the examples will include Markov Processes, as they are used in models of e.g. animal behaviour and epidemiological models and decision support in farm management.

Blog converted to Beta Blogger

During the weekend, I converted my blogs to Google's new beta Blogger. The most important result is a far more convenient handling of labelling/tagging individual posts. The conversion lead to some changes in layout. Google beta is based on a more user-friendly interface for designing the web-pages. I have tried to maintain the existing design, but I have added some new links in the sidebar

Among other new features is the possibility to create blogs with limited access, that is, only invited users can see the blogs, and the atom/RSS feeds for posts with a specific label.

Feed for News from the Danish Institute of Agricultural sciences

The home page of our Institute Danish Institute of Agricultural Sciences links to a News page. I have now made a feed for these News item.

It happens that I forget to visit the page. With other news services I simply check my netvibes home page that indicates if any new items has arrived. Unfortunately the DIAS news page does not have a feed. During the weekend I discovered the PonyFish service and generated this feed for the news. If you use Firefox 2.0 or Internet Explorer 7.0 you should simply click on the link to get started. If not, copy the link location and paste it into your favourite feed reader.

odfWeave package in R

I finally got time to read the October number of RNews. One interesting article concerned the new package odfWeave. This package gives the same functionality for using R within OpenOffice, as the package Sweave gives with LaTeX.

SWeave is used to combine text and program code for statistical analysis in a single document. It is then possible to produce a new document automatically with the results of the statistical analysis included. In our research group we use this approach more and more.

One problem with this approach (the only ?) is how to interact with people that do not use LaTeX, but Word. It seems that a large part of this problem is now solved.

I have tried it out very briefly, and it seems to be working seamlessly. The resulting document can easily be saved in Word format in OpenOffice. However, I have not tried it with the formula editor in OpenOffice

By the way, the (Danish) OpenOffice add-on Writer2LaTeX that converts OpenOffice documents to LaTeX is now (version 2.0.4) integrated in OpenOffice.

Links added to home page

I have now added some links to the right panel in my DIAS home-pages.

The first part is a search box for searches within the different domains relevant for our research group based on the Google co-op platform. This has revealed some surprising results. Some outdated pages on our old server have a high page rank and show up in the top rankings after a google search. I am now trying to update links etc.

The second part are links to different projects. The first one is a link to the Nordic-Baltic Biometric Conference 2007 to be held 6-8 June 2007, at Research Centre Foulum. I think I have discussed the other links on this blog.

Paper about Effect of zonation published

The paper about pen design for pigs mentioned previously are now available in print and with the proper volume and page information

• B.M. Damgaard, M. Studnitz, J. Nielsen, V.A. Moustsen, E. Jørgensen, and K.H. Jensen. The effects of zonation of the pen and grouping in intact litters on use of pen, immune competence and health of pigs. Livestock Science, 104(1-2):203–216, 2006.

Application for Intelligent Farrowing Pen submitted

Today we submitted our proposal concerning the intelligent farrowing pen. This grant follows our efforts in the network described on the home page Intelligent Farrowing pen.

n addition we have updated the home-page with a page describing the børsdag.

From the new page our presentations at the meetings may be downloaded. On another new page we show three video-sequences of farrowing behaviour. My first attempt at using Google Video, which is quite straight-forward.

Børsdag

On wednesday 4/10 we presented our ideas concerning the Intelligent Farrowing pen for several companies. The Børsdag was arranged by the Tech-Trans group at DIAS, in order to improve the contact between the researchers at DIAS. Our network joined this arrangement and in addition we had a separate meeting with the companies at the end of the day.

Two papers in the pipe-line

A little information about two papers in the pipeline.

The first is a paper with Ulrich Halekoh in our group as main author. The paper addresses a recurrent problem in many behavioural studies, namely the analysis of preference experiments. The paper has now been accepted for publication in Biometrical Journal

• U. Halekoh, E. Jørgensen, M. B. Jensen, L.J. Pedersen, M. Studnitz, and S. Højsgaard. Ranking of simultaneously presented choice options in animal preference experiments. Biometrical Journal, accepted, 2006.

The second paper is recently submitted to Preventive Veterinary Medicine and has S.S. Nielsen from the Department of Production Animals and Horses at the Royal Veterinary and Agricultural University in Copenhagen as main author. I have mainly collaborated with N. Toft, from the same department. The paper studies the prevalence of paratuberculosis in Danish dairy herds, based on the assumption that the observed distribution of ELISA response from diagnostic test follows a mixture distribution, with the (unknown) disease state of the animals as components of the mixture

• S.S. Nielsen, N. Toft, E. Jørgensen, and B.M. Bibby. Herd-specific sero-prevalences of paratuberculosis in 100 Danish dairy herds using Bayesian mixture models and continuous ELISA response. Preventive Veterinary Medicine, submitted, 2006.

Paper in Livestock Science available on the web

A paper about pen design for pigs are now available on the home page of the journal Livestock Science (the former Livestock Production Science).

B.M. Damgaard, M. Studnitz, J. Nielsen, V.A. Moustsen, E. Jørgensen, and K.H. Jensen. The effects of zonation of the pen and grouping in intact litters on use of pen, immune competence and health of pigs. Livestock Science, In Press, Corrected Proof:–, 2006.

Two papers now formally published

I have just checked the web-page of Applied Animal Behaviour Science, and the august number contains two papers from our project about loose housing of sows. The first one is a preference study concerning the use of pen-walls (accepted in september 2005) and the second one is the study about floor heating and piglet vitality (accepted october 2005).

Birgitte I. Damm, Vivi Moustsen, Erik Jørgensen, Lene J. Pedersen, Teresia Heiskanen, and Bjorn Forkman. Sow preferences for walls to lean against when lying down. Applied Animal Behaviour Science, 99(1-2):53–63, August 2006.

Jens Malmkvist, Lene Juul Pedersen, Birthe Marie Damgaard, Karen Thodberg, Erik Jørgensen, and Rodrigo Labouriau. Does floor heating around parturition affect the vitality of piglets born to loose housed sows?. Applied Animal Behaviour Science, 99(1-2):88–105, August 2006.

Paper now available on the internet

The paper about the use of heated floor in farrowing pen that I mentioned here is now available on the webpage of Applied Animal Behaviour Science. Just follow the link below:

L. J. Pedersen, Malmkvist, J. and Jørgensen, E. (2006) The use of a heated floor area by sows and piglets in farrowing pens. Applied Animal Behaviour Science (In press)

It is quite an extraordinary short interval from the final revision to the web-publication. However, there are quite a lot of papers waiting in the "In press" queue, but hopefully it will be formally published this year.

Paper submitted for the ISVEE meeting

Together with Niels Peter Baadsgaard from the Danish Meat Association and Vibeke Sørensen from Danish Institute for Food and Veterinary Research I have made a small paper that has been submitted for the International Symposium on Veterinary Epidemiology and Economics (ISVEE) in Cairns, Australia from 6 - 11 August 2006. The title of the paper is Evaluating A Bayesian Graphical Model For Clinical Herd-Testing and will be presented by Niels Peter at the symposium. I am not participating in the symposium. The paper is concerned with how to combine different sources of diagnostic information in order to estimate herd prevalence of different diseases, and is part of the CEPROS project about health decision support. In addition, I am also involved in another presentation at ISVEE. Nils Toft from the Royal Veterinary and Agricultural University in Copenhagen is going to present The Effect of Correlation Structures on the Properties of Diagnostic Tests for Paratuberculosis .

PigVision images to be reanalysed

Morten Birkholm, who participated in the PigVision project as an employee at Videometer, is currently participating in a course at IMM at the Danish Technical University. As a project in that course he wants to reanalyse the pig images from the experimental setup at Grønhøj. Originally the images were analysed using Active Shape Models (ASM), and now he wants to apply Active Appearence Modelling (AAM) to the images. I have sent him the approximately 1000 images with corresponding weights, and I look forward to see the results in the beginning of June. If you visit Tim Cootes' home page you can find more information about the techniques. By the way, tomorrow I am censor on master thesis about delivery strategies for slaughter pigs. Another off-spring of the PigVision project.

New version of Bayes Net home pages

Two of my home pages concerning bayesian networks have been relocated/redesigned. The first is examples of use of bayesian network in animal production, and the second is the web version (in Danish) of the article in Naturens Verden: Bedre Beslutninger med Bayesianske Netværk, which I coauthored with Steffen L. Lauritzen.

Paper accepted

Our paper has been accepted about the use of heated floor in farrowing pen from the project about loose housing of farrowing sows. The first version was returned with a request for minor revisions, which we have made. The formal citation is:

L. J. Pedersen, Malmkvist, J. and Jørgensen, E. (2006) The use of a heated floor area by sows and piglets in farrowing pens. Applied Animal Behaviour Science (accepted)

I had to make some recalculations concerning the diurnal rythm in behaviour when farrowing approaches. It was a bit tricky because we considered time on two different scales, the chronological scale (time of day) and the time related to time until/since birth of first pig.

Network for Intelligent Farrowing Pen

Our application about the Intelligent farrowing pen of the future has been granted. The purpose is to establish a network between researchers, IT-companies and Housing companies. Lene Juul Pedersen is the project leader.