Projects address a problem of interest, often focused on graduate research. The final report and presentation informs the class about your project. The report should not exceed 8 pages of text and graphs. Any code goes in an Appendix.

Structure of the report:

We will circulate the final reports to reviewers (each other) for feedback. Grading will be based on the peer reviews.

Setting up the model will involve 1) constructing a graph of the model, 2) writing down functional forms, and 3) simulating data. The graph (step 1) provides the road map used to construct the full model and the algorithms needed for computation. The functional forms (step 2) embody the processes represented by arrows in the graph and represent relationships among model elements. Data simulations (step 3) can be used to assess whether the algorithm is doing what we think it should be doing. Exploratory data analysis (EDA) should be provided to give the reader a ‘feel’ for pattern in the data.

This assignment consists of two parts, your project summary and your review of two other projects in your group.

Your project summary (next week)

(3 paragraphs, 1 graph, pseudocode for simulation)

Note that there must be a real question, in the sense that more than one answer is possible. If only one answer is possible, it is not really a question.

Here is an example:

Data summary–Data consist of nutrient concentrations in extracts from incubated soil cores, obtained monthly from each of 15 plots, 10 replicates per plot and month. Observations are positive continuous variables.

Model summary–Consider a soil nutrient pool that is supplemented by litter fall in autumn and releases nutrients as litter degrades in mass and quality until renewed the following autumn. The model describes increase in mineralization rate as temperatures and moisture increase in the spring and then decline with the decrease in labile organic matter. State variables are litter mass and litter quality. Predictors of mineralization (nutrient release) are temperature, moisture, litter mass, and litter quality. The response variable is nutrient concentration. Parameters describe how each predictor affects litter decay and nutrient release. There are random plot effects.

Include subscripts to indicate how these relationships are treated in the model.