If you'd like to view your exam paper, you can drop into my office (608/K17) between 12noon-1pm on Thur 6 July.
The exam and final results are now available, and you can check them from the course website.
The exam marks that you see on the system have been scaled. I used this scaled mark together with your coursework marks (coming from adding the assignment and project marks) to obtain the final mark. The only mark that I adjusted is the exam mark, all the other marks have not been adjusted.
The final mark is computed using a harmonic mean formula and you can find the formula in the course outline. If your final mark is 50 or more, then you have passed the course. Otherwise, you have unfortunately not made it.
The CSE supplementary policy says that if you have attended the exam, you will not be granted a supplementary exam. You can find this policy at:
https://www.engineering.unsw.edu.au/computer-scien...
I will make a time slot available on Thur 6 July for viewing your exam paper without an appointment. I'll make an announcement when I know the time.
It's been a pleasure teaching you all. I hope you enjoyed the course and have learnt something useful. All the best.
The marks for the project is now available. You can check your mark via the course web site. The class did quite well for the project with an average mark of 15.2. The project was marked according to the requirements set out in Section 4.2 of the project specifications. Some common places where marks were lost are:
1. Some students did not provide evidence that they had verified their simulation code. The specifications ask you to use test cases.
2. Some students did not justify the number of replications they chose. You need to connect the number of replications to the confidence intervals you are getting.
3. Some students did not compute confidence intervals.
4. Some students made their choice of the number of servers based on the mean response time. This is not statistically sound. The decision has to be made by comparing confidence interval. You can either use the visual test to compare the confidence interval or you can use the paired t-test.
5. Continuing on Point #4 above. Some students used mean response time to narrow down the number of servers to a subset (e.g. 6, 7 or 8 servers) and then use confidence intervals to determine the exact number of servers. The first part which is based on mean response time is not statistically sound.