diff --git a/Index.html b/Index.html index 570a208..51c3837 100644 --- a/Index.html +++ b/Index.html @@ -1456,9 +1456,10 @@

Lecture 6: Building Decision Models

steps in the process, i.e. starting with a skateboard rather than a car part as in this example from MetaLab.

-
- -

‘Skateboard, Bike, Car’

+
+ +
‘Skateboard, Bike, Car’

Before you start developing the decision model (an R function), open @@ -1661,7 +1662,7 @@

Plot the impact pathway

into code for estimating profit distributions.

We use the mermaid function from the DiagrammeR library to create the graphical impact pathway -(Iannone 2022).

+(Iannone 2023).

Run the code below and then add an additional cost to the graph called Management cost (try to do this on your own but feel free to look to the solution to see one way to do @@ -2796,11 +2797,12 @@

The principle of Bayesian statistics

parameter value plotted on the x-axis. The total area under the curve always adds up to 1 (exactly some parameter value must always be true).

-
- -

Source: + +

Source: https://flexbooks.ck12.org/cbook/ck-12-probability-and-statistics-concepts/section/7.5/primary/lesson/probability-density-function-pst/.

+class="uri">https://flexbooks.ck12.org/cbook/ck-12-probability-and-statistics-concepts/section/7.5/primary/lesson/probability-density-function-pst/.

So how do we calculate this probability? Unfortunately, it’s usually not as straightforward as it seems. While we can usually calculate the @@ -2821,9 +2823,10 @@

MCMC - Markov Chain Monte Carlo

Github page at https://michael-franke.github.io/intro-data-analysis/Ch-03-03-estimation-algorithms.html. Here are some parts of it:

-
- -

Image by brgfx on Freepik

+
+ +
Image by brgfx on Freepik

How the Apples Get to the Trees

Every year in spring, Mother Nature sends out her children to @@ -2867,15 +2870,16 @@

MCMC - Markov Chain Monte Carlo

Result2/Result1. Then, values are drawn randomly again, and so on. In the following figure, successful jumps are represented by blue arrows and rejected jumps by green arrows.

-
- -

Visualization of an MCMC simulation. You have to +

+ +
Visualization of an MCMC simulation. You have to imagine that you cannot see the salmon-colored parameter landscape, which represents the values for Likelihood * Prior, but can only calculate the values of individual points. Further explanations are given in the text. Source: https://www.turing.ac.uk/research/research-projects/adaptive-multilevel-mcmc-sampling

+class="uri">https://www.turing.ac.uk/research/research-projects/adaptive-multilevel-mcmc-sampling

If this process is continued long enough, the distribution of the blue dots approaches the posterior probability distribution of the @@ -2993,10 +2997,11 @@

Frequentist versus Bayesian

The easier interpretability of Bayesian analyses is best illustrated by the following figure:

-
- -

Source: Korner-Nievergelt und Hüppop (2016). Five -possible results of estimating an effect, such as the difference in +

+ +
Source: Korner-Nievergelt und Hüppop (2016). +Five possible results of estimating an effect, such as the difference in yield with different fertilization. The dots indicate the estimated differences, the vertical bars indicate the 95% uncertainty intervals (confidence interval or credible interval). The results of the @@ -3004,7 +3009,7 @@

Frequentist versus Bayesian

second row shows the posterior probability for the hypothesis that the effect is “economically relevant”. The background color distinguishes “economically relevant” (orange) from “economically not relevant” (light -blue) effect sizes schematically.

+blue) effect sizes schematically.

An uncritical frequentist approach would result in: “The effect in group/treatment A and E is significant, so we will use the tested @@ -3367,7 +3372,7 @@

Lecture 10: Profile of a Decision Analyst

Whitney or the course tutor.

Those who work in Decision Analysis must be able to do a number of diverse jobs from the facilitation and integration of ideas into models -and also in programming these models. In teh following short lecture you +and also in programming these models. In the following short lecture you will learn about the skills that are particularly important for integration of knowledge, facilitation of knowledge gathering processes and for programming decision models. Please @@ -4008,22 +4013,22 @@

References

@@ -4108,21 +4113,21 @@

References

@@ -4142,15 +4147,15 @@

References

@@ -4171,17 +4176,17 @@

References

@@ -4233,21 +4238,21 @@

References

@@ -4298,20 +4303,20 @@

References

@@ -4330,27 +4335,27 @@

References

@@ -4372,27 +4377,27 @@

References

@@ -4413,31 +4418,31 @@

References

@@ -4459,21 +4464,21 @@

References

@@ -4527,21 +4532,21 @@

References

@@ -4597,22 +4602,22 @@

References

@@ -4668,17 +4673,17 @@

References

@@ -4724,22 +4729,22 @@

References

@@ -4790,14 +4795,14 @@

References

@@ -6296,24 +6301,24 @@

References

@@ -6335,26 +6340,26 @@

References

@@ -6420,17 +6425,17 @@

References

@@ -6612,21 +6617,21 @@

References

@@ -6647,17 +6652,17 @@

References

@@ -6707,16 +6712,16 @@

References

@@ -6762,16 +6767,16 @@

References

@@ -8940,16 +8945,16 @@

References

@@ -8999,22 +9004,22 @@

References

@@ -9075,22 +9080,22 @@

References

@@ -10258,44 +10263,44 @@

References

@@ -10316,50 +10321,50 @@

References

@@ -11396,12 +11401,12 @@

References

@@ -11468,7 +11473,7 @@

References

Second Edition. Hoboken, New Jersey: John Wiley & Sons.
-Iannone, Richard. 2022. DiagrammeR: Graph/Network +Iannone, Richard. 2023. DiagrammeR: Graph/Network Visualization. https://github.com/rich-iannone/DiagrammeR.
diff --git a/Lecture_10_Analyst_Profile.Rmd b/Lecture_10_Analyst_Profile.Rmd index d46f4df..c8cf487 100644 --- a/Lecture_10_Analyst_Profile.Rmd +++ b/Lecture_10_Analyst_Profile.Rmd @@ -5,7 +5,7 @@ Welcome to lecture 10 of **Decision Analysis and Forecasting for Agricultural Development**. Feel free to bring up any questions or concerns in the Slack or to [Dr. Cory Whitney](mailto:cory.whitney@uni-bonn.de?subject=[Lecture_6]%20Decision%20Analysis%20Lecture) or the course tutor. -Those who work in Decision Analysis must be able to do a number of diverse jobs from the facilitation and integration of ideas into models and also in programming these models. In teh following short lecture you will learn about the skills that are particularly important for *integration* of knowledge, *facilitation* of knowledge gathering processes and for *programming* decision models. Please watch the video and answer the multiple choice questions below. +Those who work in Decision Analysis must be able to do a number of diverse jobs from the facilitation and integration of ideas into models and also in programming these models. In the following short lecture you will learn about the skills that are particularly important for *integration* of knowledge, *facilitation* of knowledge gathering processes and for *programming* decision models. Please watch the video and answer the multiple choice questions below. diff --git a/bib/packages.bib b/bib/packages.bib index 277bec1..b3aec7e 100644 --- a/bib/packages.bib +++ b/bib/packages.bib @@ -26,8 +26,8 @@ @Manual{R-decisionSupport @Manual{R-DiagrammeR, title = {DiagrammeR: Graph/Network Visualization}, author = {Richard Iannone}, - year = {2022}, - note = {R package version 1.0.9}, + year = {2023}, + note = {R package version 1.0.10}, url = {https://github.com/rich-iannone/DiagrammeR}, } @@ -92,7 +92,7 @@ @Manual{R-igraph title = {igraph: Network Analysis and Visualization}, author = {Gábor Csárdi and Tamás Nepusz and Vincent Traag and Szabolcs Horvát and Fabio Zanini and Daniel Noom and Kirill Müller}, year = {2023}, - note = {R package version 1.4.2}, + note = {R package version 1.4.3}, url = {https://CRAN.R-project.org/package=igraph}, } @@ -100,7 +100,7 @@ @Manual{R-knitr title = {knitr: A General-Purpose Package for Dynamic Report Generation in R}, author = {Yihui Xie}, year = {2023}, - note = {R package version 1.42}, + note = {R package version 1.43}, url = {https://yihui.org/knitr/}, } @@ -156,7 +156,7 @@ @Manual{R-rmarkdown title = {rmarkdown: Dynamic Documents for R}, author = {JJ Allaire and Yihui Xie and Christophe Dervieux and Jonathan McPherson and Javier Luraschi and Kevin Ushey and Aron Atkins and Hadley Wickham and Joe Cheng and Winston Chang and Richard Iannone}, year = {2023}, - note = {R package version 2.21}, + note = {R package version 2.22}, url = {https://CRAN.R-project.org/package=rmarkdown}, }