[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Project part b grading remarks from the TA



I will return the projects today.

Rao
[Nov  9, 2005]

==========
Stats:

Points are given as following:

A&H report: 15
pageRank Report: 15
output: 10
coding: 20
GUI(extra credit): 5
Total: 55 + 5


The following are the stats (extra credit not included):

avg: 49.7
max: 60
min: 27
stdev: 10.6

graduate avg: 49.7
graduate max: 60
graduate min: 27
graduate stdev: 11.5

undergrad avg: 49.5
undergrad max: 59
undergrad min: 40
undergrad stdev: 9.4

================

The grading is based on roughly the following points:
1. whether the implementation gives reasonable answers for the testing queries
2. whether the coding takes advantage of the index structure and link 
extraction model we gave to compute A/H and pagerank in a reasonably efficient 
fashion
3. the analysis of the algorithms and the implementation, say
   1) time and space complexity analysis
   2) the relevancy of the answers given by different models.
   3) the influence of the root set size of A/H computaion on the search 
results and the convergence of the computation
   4) whether some actual expriments were done to test the effect of varying 
the weight to combine vector similarity and pagerank value.
   5) experiment on different damping factor in pagerank computation and the 
analysis of its impact on the search result and convergence speed etc.

Overall I think most people did a good job in this project. Most of them have 
nice coding and reasonable search result and analysed the various topics. The 
3 or 4 not so good ones are mostly having basic coding problems: they did not 
implement the algorithms right (at least 2 of them are caused by the leftover 
problems of the first project) and thus their output and analysis are 
problematic too.


Thanks,
Jianchun