15.062  Data Mining: Finding the Data and Models that Create Value

Dear All,

I've just posted the datasets for homework 3 and today's recitation slides.

http://stellar.mit.edu/S/course/15/sp09/15.062/courseMaterial/topics/topic6/resource/Homework_3_Datasets/Homework_3_Datasets.zip

Cheers,

- mitu

Dear All,

Homework 3 has been posted and it is due on March 11.  Also, if you haven't emailed me your project proposal yet, please do so as soon as possible.

Regards,

- mitu

Dear All,

As part of the German Credit Case, please specify all the data manipulations that you perform, if any.  Please list these by the modelling algorithm. 

I have graded the HW1 and returned most of them.  I still have a few with me.  The score distribution was: Average 39, Median 41, Std. Dev 7, Max 46.

Regards,

- mitu

Dear All,

There's some confusion about the German Credit case.  Please answer the following questions:

1) Text Question 1

2) Modify Text Question 2 to the question posed in the assignment.

3) Report confusion matrix and the cost/gain matrix for the validation set.  Calculate netprofit for each of the models.  Let us know which model do you think is the best one.

4) Suggest any improvements etc to the base models.

Please let me know if you have any questions.

Regards,

- mitu

Dear All,

I've posted recitation slides for today.

Cheers,

- mitu