>> Applied Statistics
Assignment - Decision Trees
The goal of this is to construct a decision tree to model and evaluate decision options using that model.
Create a decision tree for analyzing a decision of your choice. Make sure the following issues are addressed:
- Estimate probabilities and costs (values) based on data, literature or by interviewing an expert
- Calculate expected value (cost) for each decision using one of the methods described in the lecture. Make sure values are non-trivial (i.e., if costs for one decision are always higher than for other decision, then there is no point to construct tree)
- Perform a single-parameter sensitivity analysis (Note that single parameter means that one parameter is changed, but many parameters need to be tested)
- The decision tree should have at least four levels (including the decision node)
- Present discussion and conclusions. What does the model tell you about your decision? Is the model valid? It is sensitive to how it was constructed?
Your decision tree can model either a single decision of a sequence of decisions. Prepare narrated Powerpoint that reports on results of your project.