Download File - Amy & Steve CSCI 5980 GIS - Home

raviolirookeryΒιοτεχνολογία

2 Οκτ 2013 (πριν από 3 χρόνια και 10 μήνες)

74 εμφανίσεις

CSCI5980

Computational Geo Informatics

Amy Casuscelli



Steve Hawkins


Csci 5980 Fall 2010 Final Critique Form

Please use the following questions to critique each group:

Group 6
: Temporal Association Rule Mining


Problem Motivation:

Did the speaker motivate this problem? Why is the problem
interesting? What applications can this problem be applied to? List at least two
applications this problem can be applied to other than what was presented. Briefly
justify your answer. (50 word lim
it)


The initial example of the grocery store analysis engaged us to understand the topic.
We immediately thought this would be useful in identify patterns of fraud in financial
systems. They mentioned that the problem could be applied in:

Retail Data An
alysis

Web Usage Mining

Intrusion Detection

Bioinformatics


I think that there may be application in the OAAM context to examine and improve the
fraud rules based on a statistical analysis of the set of data over time. The temporal
aspects of that data wo
uld play an important role.


Relates to Chapter 10

Transaction Time

Valid Time


Apriori Algorthim


Lim
itation, find the frequent item
sets in the transaction database,
hard
item

find those that are infrequent.


Time Instants or Time Intervals


Problem
Statement:

What problem did the presenter attempt to solve? If applicable,
what were the input, output, objective, and constraints for this problem? Briefly justify
your answer. (50 word limit)


All temporally frequent itemsets along with their maximal val
id time intervals.

(from
their slide) Which we believe that they are attempting to determine the relationships
between data for a maximum time interval. Their algorithm would find that optimal
maximum value.


Hardness:

What is the main challenge in this
work? Classify these challenges into
computational, spatial, other, etc. Briefly justify your answer. (50 word limit)


CSCI5980

Computational Geo Informatics

Amy Casuscelli



Steve Hawkins


This work has a computational complexity but I suspect it is of interest to application
owners to attempt to solve.


Proposed Solution:

H
ow did the presenter solve the problem that was described? What
were the key ideas behind the proposed solution? What was innovative about their
solution? List at least three suggestions on how their proposed solution may be
improved. Briefly justify your
answer. (50 word limit)


The discussion was around the algorithm to limit the windows in which could be
analyzed. They revealed a set of rules that could be used to limit or expand the window
for analysis. One thing we did not discuss was if the team had

implemented the
algorithm and run any test cases. We have not suggestions for improvement at this
time.


Related Work:

Did the speaker provide a sufficient survey of related work? What were
the main limitations in the related work? Do you feel that eithe
r the problem or
proposed approach is novel? Briefly justify your answer. (50 word limit)


The team provided some good examples of other algorithms, we had not heard of the
Apriori algorithm before. Limitations were infrequent items were hard to analyze a
nd
not much
representation

of temporal data. Other related work mentioned the
definition of valid time approach was complicated.


Validation:

What were the validation methodologies used in this work? Was the
proposed approach compared with any state
-
of
-
th
e
-
art or naive approaches? Were
there any surprises in the validation results relative to your expectations? Briefly justify
your answer. (50 word limit)


The team discussed their algorithm but validation was not discussed.


Presentation Critique:



Do you

think the speaker did a good job motivating and defining the problem clearly for
someone like you who may be a "non
-
expert" on the topic? Why or why not? (50 word
limit)


Given the time limit our team feels it would need further discussion to understand
the
core problem they are trying to solve.


Cite specific use cases would be a great addition.

Did the speaker emphasize the central message (e.g. did they follow:
"Tell the audience
what you're goin
g to say, say it; then tell them what you've said."
) (50 word limit)


CSCI5980

Computational Geo Informatics

Amy Casuscelli



Steve Hawkins


I think a recap summary at the end would have been helpful in relating the project to a
real world need.


Did the speaker relate to the audience? What did the speaker do to establish a r
apport
with the audience before delivering the key message? (50 word limit)


They were very conscientious in engaging the audience to ensure we were keeping up.
Looked at the audience while speaking and were enthusiastic. The example of the
grocery store

set a good baseline of understanding.


What did the speaker do to engage the audience? e.g. Did the speaker ask short
questions to either motivate the upcoming topic or assess the background/thinking of
the audience? (50 word limit)


see above.


Did you
understand most of the talk (or 75%, 50%, 25%, 5%, 0%)? Did the speaker
inspire you to want to learn more about what he/she is doing? If you had to rate the
presentation check
-
, check or check+, what score would you give? Briefly justify your
answer. (50 w
ord limit)


We felt we could follow the speaker’s ideas. The concept seems like it would be useful
in fraud analysis which our presentation was on. We see some potential overlap.
Check+. Good work.