Parametric statistics are based on the fitting of hypothesized mono- or multivariate linear or non-linear models relating dependent and independent variables. The hypothesized relationship and its associated parameter estimates are commonly used for modelling natural resource availability and forecasting their development as a basis for sustainable management. The correct use of statistical methods, models and approaches plays a pivotal role in achieving appropriate and unbiased data summaries and statistical inference and hence providing reliable projections of present and future resource availability.
The course provides the students with the tools to select appropriate statistical models, to estimate their parameters, to evaluate derived models, to make statistical inference from their findings, and to apply models in simulation and assessment of forest and other natural resources.
- Selecting proper statistical methods and suitable models for different types of data and data sampling
- Understanding model assumptions and their implications
- Different model fitting approaches
- Statistical inference
- Post-calibration model evaluation
- Model simulations and their validity
|Week 3 and 4:||Pre-course assignments including reading of course portifolio and selection of personal dataset|
|Feb 1st:||Course introduction, model selection|
|Feb 2nd:||Model estimation (linear and non-linear models)|
|Feb 3rd:||Post-hoc analyses, fit statistics, cross validation|
|Feb 4th:||Model inference, resource assessment and simulation|
|Feb 5th:||Course round up, report supervision|
|Week 6:||Preparing reports for course evaluation|
- describe principles for selecting appropriate statistical models and calibration procedures for different types of data
- understand the relationship between statistical analysis and development of models
- show overview of model types used to describe relationships and to model natural resource dynamics
- apply statistical principles and methods in typical natural resource modelling situations
- select suitable model formulations for modelling particular relationships
- apply suitable methods for assessing the quality of predictions
- apply general modelling and forecasting principles, involving typical variables from forest and nature
- discuss the relevance, reliability and interpretation of empirical statistics
Prior to the course, students will read the course portifolio, mainly consisting of scientific articles related to the course subject. Further, the students will choose a relevant dataset to analyze during the seminar in week 5. If a student lack a relevant dataset, the course responsible will provide one. After the course, the students will prepare a written report on the analyses conducted and their results. A satisfactury evaluation of the report is necessary to pass the course.
The students will be evaluated based on a report describing the analyses conducted during and after the seminar and their reults.
Students are expected to work independently with their own material prior to and after the course seminar. During the seminar week, students will receive lectures detailing common approaches in statistical modelling and working with small examples. After lectures, students will work with their own datasets, receiving supervision by the course team.
- 20 hours seminar
- 20 hours lecture
- 80 hours independent work
The students are required to have basic statistical competences in statistics and regression analysis.
Admission for NOVA courses is handled by the course organiser/ the NOVA member institution organising the course. Please see the links in the margin for more information.
- Accommodation will be at the forestry college in Nødebo. Accommodation costs are EUR 250 for the full week.
- For non-NOVA and non-BOVA students, there is a course fee of EUR 200.