Exercise 3: Clustering with K-Means and SOM methods.
Submission:
1. Deadline March 24, 2015, Tuesday, before 11:59p.
2. Electronic submission is required. Send the file to isqs6347@gmail.com with subject title: “ISQS 6347 EX3, <your name>”. Without clearly specifying the contents in the title may result in missing grading.
3. File naming convention: ISQS6347-EX3-2015-<Team Letter>-<your last name>. For example: ISQS6347-EX3-2015-R-Lin
Late Submission Policy
1) Late submission without notification and grant will be deduced 10% of the total.
2) Highly suggested that students start working on the Exercise on March 11, Wednesday before the spring break, because if any server problem occurs students can get help in regular office hours.
3) If a student could not complete the exercise in time because of computing system’s problem, he/she needs to email to Anagha for the extended submission. Up to 2 days are fine after the approval of Anagha as authorized.
References:
1. CCWEB_TKIT.pdf
2. ADMT_001.pdf (in the shared network drive under ~\Textbooks\Others\)
-----------------------------------------------------------------
1. Dataset.
The dataset is DMAIL, in directory \\bamsasfs\data\DATA_CCWEB\. The explanations of the dataset can be found in textbook CCWEB_TKIT. Check page 4-9 of the book (See Figure 1). You are requested to do clustering with this dataset (Figure 2) using both k-means and SOM methods available in SAS Enterprise Miner, and compare the results of the clustering model with those from the classification models in the CCWEB_TKIT.pdf (or in Homework #2 if you have done it). For information about SOM/Kohonen model check Section 7.2 in SAS courses notes ADMT_001.pdf and slides of DM0-PatternDiscovery.pptx.
Figure 1. The dataset
Figure 2. The model
2. Node configurations
1) Input Source node - DMAIL:
a. Set ResponseFlag and TotalSpent as “Rejected”
b. Make sure the type of ProspectID is ID
2) Cluster node: Standardization, 8 clusters
3)
