The Projects
Scroll down to check the general guidelines and reporting criteria follwing this section.
Project 1: Special Discrete Probability Distributions
- Due: Saturday April 2, 2005
- Grading criteria:
- Quality of experimentation.
- Quality of discussion and interpretations of shapes.
- Quality of examples for each shape (basically, generic distribution textbook examples vs. well thought-out examples to explain the probability trend exhibited by each shape).
Project 2: Simulation of Generalized Bernoulli Trials
- Due: Sunday May 1, 2005
- The detailed criteria:
- Discussion and (summary) presentation of results.
- Comparison with theory.
- Insights: trends, analysis and interpretation (I need to see some probabilistic reasoning which reflects STAT612 knowledge).
- Conclusions.
- Report presentation: structure and organization.
- Grade guide:
- A: top job, best probabilistic analysis, best link to theory, best reporting.
- B+: mostly good experimentation, very weak probabilistic analysis.
- B: need to work harder.
Project 3: Simulation of Compound Poisson Proces
- Due: Tuesday May 31, 2005
- Please read general criteria below and the criteria for Project 2 carefully.
General Guidelines
Simulation
- The simulation study should generally proceed in the following pattern: explore behaviors (careful design or selection of cases), compute & summarize results, discuss findings, show trends, compare with theory (or previous work), and draw conclusions.
- It's OK to briefly repeat the problem statement, and to include selected textbook definitions for background and context. However, this should not be at the expense of the time spent thinking about the experiment, the results, the conclusions, and studying the book for answers.
Results Summary
Conventionally, there are 3 ways to summarize in scientific communication: tables, graphs, and math formulas. In some cases, diagrams may be appropriate.
Project Report Criteria
The basic work of programming the simulation will not receive any weight as this is the main skill of CS students. Everybody is expected to do a good job. It is a pre- requisite and is not related to performance in STAT612.
The exception is programming in Maple. Demonstrating skill with Maple will be rewarded appropriately.
Otherwise there are 3 broad criteria:
- Designing the experiment: selecting appropriate parameters and test cases.
- Reporting: a skill under development in STAT 612 therefore relatively little weight will be assigned in projects 1 and 2. In project 3, however, I do expect to see clear progress. Your report should fully reflect whatever guidelines I publish or comments I make about reporting. A significant weight will be assigned in project 3.
- STAT 612 knowledge: a measure of the quality of study and use of the textbook. The majority of weight is always assigned to this category.