MGS 8040 Data Mining Project
Written Report
- Introduction
– what is the goal of the project?
- Data
- Source,
variables (put data dictionary in appendix)
- Sample
Size
- Data
Preparation (aggregation, variable creation, data cleaning)
- Methodology
- Example
of crosstabs, dummy creation
- Regression
(discrim/logistic)
- Results
- Show
scorecard (results of final regression in plain English)
- (put
actual regression results – the 1st and the last regression
only in appendix)
- KS
test results for training and validation samples.
- Discussion
of improvement in performance that client will get from your model.
- Implementation
- Discuss
how the client should implement your model – what cutoff scores you
recommend, what strategies go with the cutoffs.
- Monitoring
Reports
- Explain
how the client should monitor the performance of the model after
implementation.
- Project
flow diagram: Include a diagram that shows all the files created, the
programs used to create them, and other actions taken using those files.
This is for your benefit rather than the client’s.
Oral Presentation
Since we are all following similar methodology, there is no
need to discuss that in your presentation. Present only the results – the
scorecard and the KS tests. Ten minutes should be sufficient.
Sample Scorecard
|
|
Variable
|
Intervals
|
Points
|
|
0
|
Intercept
(All applicants start with this score)
|
|
459
|
|
1
|
Customer Age
|
Missing
18-25
26-40
41-50
51+
|
0
-25
0
31
64
|
|
2
|
Number of Revolving Trades
|
No Record
Inqs only
Pub Rec only
Inqs and Pub rec
only
Missing
0-3
4-9
10-13
14-20
21+
|
0
0
0
0
0
-44
0
35
55
35
|
|
3
|
Retail Value
|
$ 0- 1000
1001 – 5000
Etc…
|
|
|
4
|
And so on …
|
|
|