SASK22001U Quantitative Methods in Herd Management

Volume 2024/2025
Education

MSc Programme in Animal Science - restictive elective

Content

This course is for people seeking to improve their abilities to assemble and use data of animal performance to enhance rational, quantitative decision making in animal production. It is a computer intensive course, where the students will need to install specific programs (R and R studio) as well as specific R packages on their computers. We will be using real-world data sets collected in commercial herds and in research herds.
 

The students will write and submit weekly reports. In these reports, they will demonstrate their abilities to use the methods and explain the concepts introduced in a given week. When making these reports, the students will work together in groups of 2-4 people. 

The necessary mathematical and statistical preconditions will be introduced at appropriate times, relative to the topics being covered in a given week. This is done in order to make sure that all students have a basic understanding of topics like vector and matrix operations (i.e. linear algebra), random numbers, Bayes’ theorem, distributions (including multivariate and conditional), etc. 

Next, the course provides a comprehensive introduction to quantitative herd management by combining:

  • Theory
  • Computer applications in order to illustrate theory
  • Presentations of practical livestock management applications developed in research
  • Presentations from relevant guest lecturers 


The techniques covered in this course have general applications in diverse animal production systems. These techniques are applied to an array of management decisions including culling, breeding and mating; feed allocation and slaughter timing; and medical treatment.

 

The course will cover the following specific topics:

  •  Linear algebra and R-programming
  •  Monitoring of time series data using classical methods (i.e. moving average, exponentially weighted moving average)
  •  Monitoring of time series data using state space models
  •  Random forests
  •  Artificial neural networks
  •  Markov decision processes
  •  Monte Carlo Simulation

 

Learning Outcome

After attending the course, students should be able to participate in the development and evaluation of new tools for management and control, while taking biological variation and observation uncertainty into account.

After completing the course, the students should be able to:

Knowledge:

  • Describe the methods taught in the course.
  • Explain the common pitfalls, and how to avoid them, when using specific methods taught in the course. 
  • Explain the limitations and strengths of the methods in relation to various herd management problems.
  • Give an overview of typical application areas of the methods.


Skills:

  • Write and apply their own software tools in R.
  • Apply the other software tools used in the course.
  • Train and validate machine learning methods for classification and regression tasks.
  • Construct models to be used for monitoring and decision support in animal production.


Competencies:

  • Combine the different methods in meaningful ways in response to a given herd management problem.
  • Transfer methods to other herd management problems than those discussed in the course.
  • Evaluate the validity of machine learning models in terms of architecture, training, and validation strategies.  
  • Evaluate methods, models, and software tools for herd management.
  • Interpret results produced by models and software tools.

Kristensen, A.R., E. Jørgensen and N. Toft. 2010. Herd Management Science I. Basic concepts. 2010 Edition, University of Copenhagen, Faculty of Life Sciences.

Kristensen, A.R., E. Jørgensen and N. Toft. 2010. Herd Management Science II. Advanced topics. 2010 Edition, University of Copenhagen, Faculty of Life Sciences

Kristensen, A.R. 2010. Herd Management Science. Exercises and supplementary reading. 2010 edition.
 

Additional scientific papers and other materials, which will be relevant to the topic of a given week. 

A Bachelor's degree in Animal Science (Danish: Husdyrvidenskab) or similar.
It will be an advantage for the student to have taken one or more of the following courses (or similiar) before taking this course:

LMAB10066U Matematik og databehandling
LMAB10069U Statistisk dataanalyse 1
Lectures, theoretical exercises, computer exercises, and report writing.

In general, each week will cover one specific topic. The students will make a mandatory report relating to the topic of each week. The students can work on the reports in groups of 2-4 people.

The submitted reports will be assessed regularly. If a report is not accepted in the first attempt (including if the report is not submitted on time), the students have one more chance to submit a corrected report, utilizing the feedback provided in the assessment.

In connection with lectures, the students are expected to participate actively in discussions. Throughout the course, the application and limitations of the methods taught will be illustrated by larger examples presented by relevant guest lecturers.

Lectures will be supported by theoretical problem solving exercises, and all methods presented will be supported by computer exercises where the application of the methods is illustrated. The reports will consist of answers to selected exercises, evaluation of methods, and one or more implementations of methods used to solve management problems.

Evaluation model: Survey-based model
  • Category
  • Hours
  • Lectures
  • 42
  • Preparation
  • 108
  • Theory exercises
  • 42
  • Project work
  • 200
  • Guidance
  • 19
  • Exam
  • 1
  • Total
  • 412
Continuous feedback during the course of the semester
Credit
15 ECTS
Type of assessment
Oral examination, 25 minutes
Type of assessment details
An individual oral examination is held at the end of the course.

The student randomly draws an exam question related to one of the topics taught in the course.

The student then randomly draws one out of 5 reports. If the student has succesfully completed 6 or 7 reports, the student can choose one or two reports, respectively, to omit from the random draw.

The student has 30 minutes of preparation time, including the drawing of the exam question and report. This is followed by 25 minutes of examination, followed by 5 minutes of deliberation and grading.

Students in the same group cannot participate in their fellow student’s individual oral exam unless they have already had their own individual oral exam.
Exam registration requirements

A precondition for attending exam is that at least 5 of 7 mandatory reports have been submitted and approved.

If the student has succesfully completed 6 or 7 reports, the student can choose one or two reports, respectively, to omit from the exam. 

The student may still draw an exam question related to the topic of the omitted reports, so it will be advantageous to the student to complete all reports, even though only 5 completed reports are strictly required. 

Aid
All aids allowed
Marking scale
7-point grading scale
Censorship form
No external censorship
More than one internal examiner.
Criteria for exam assesment

To achieve the maximum grade of 12, the student must be able to:
 

Knowledge:

  • Describe the methods taught in the course.
  • Explain the common pitfalls, and how to avoid them, when using specific methods taught in the cause. 
  • Explain the limitations and strengths of the methods in relation to herd management problems.
  • Give an overview of typical application areas of the methods.

 

Skills:

(Only tested in mandatory reports)
 

Competencies:

  • Evaluate the utility and validity of the methods, models, and software tools for herd management problems
  • Interpret results produced by models and software tools
  • Describe meaningful combinations of various methods for specific herd management problems