Continuous Assessment Overview
Module 7 KAN Min-Yen

Survey Paper
Grading Metric:
Average:
Std. Dev:
Disputes on
grade should be brought up by the end of this week

Proposals
Need a clear formulation of the problem:
Scope of the problem
Limit what you are going to examine
Be precise about what you are going to tackle
Motivation - reason why applicable to DL
A reasonably detailed idea of how you will approach the problem
How would one evaluate your project?

Timeline for Final Project
24 Sep: Today
2 week interval: have a clear idea what your project is going to be
15 – 22 Oct: lecture presentation (done in pairs, sign up during break)
2 week interval: have finished the bulk of the programming, and thought about how to evaluate your work
13 Nov (Thursday, from 1-5 pm): final presentation

Lecture presentation groups
HENDRA SETIAWAN & EDWARD WIJAYA (Mixed-media Mining in DL)
LIN LI & WONG SWEE SEONG (Document Clustering for DLs)
HUANG WENDONG & WANG GANG (Music & Speech in Digital Libraries)
LI YINGGUANG & VOROBIEV ARTEM (Peer to Peer in DLs)
GUO SHUQIAO & YANG HUI (Web/Hypertext Information Seeking)
MASLENNIKOV MSTISLAV & CHAN YEE SENG (Spatial and Temporal DLs)
QIU LONG & TOK WEE HYONG (Intelligent Agents in DLs)
CHEN XI & WANG XIAOHANG (Metadata Extraction and Indexing)

Project grading dimensions
Grade will depend more on quality, ideas, crisp results than number of hours spent
Creativity in defining the problem to be investigated
Quality of the methods used investigating it
Thoroughness in justifying your design decisions
Quality of your write-up
Reporting methods, results, discussion, etc.
Quality of evaluation of your system
You will not be penalized if your system performs poorly

Quality
Try to make it something that shows something interesting, not just an exercise in programming
But it can be a small something – be focused and not too ambitious

What should be in the report?
A clear research question or application, and hypotheses about a good approach to it
A clear and complete discussion of the theory / algorithms / method used
A high-level description of the implementation
Testing, and discussion of results
Clear graphs, tables, experimental comparisons
A discussion of alternatives and their performance
Brief but adequate discussion of related work

Rough grading guidelines
Idea’s motivation, originality, scoping: 25%
A chance for you to refine your proposal
Experiment methods, implementation: 25%
Evaluation methods, results and post-analysis: 25%
Quality of write-up
Including adequate comparison to prior work, proper citing: 25%
This part should be easy since you have done the survey paper

Outline for today…
Bibliometrics
Citation typing and visualization
Pagerank
HITS