1The Hang Seng University of Hong Kong
Dept. of Supply Chain and Information Management
MSc in Global Supply Chain Management
SCM6005: Class work 03
Question 1
Suppose you are a boss of a coffee shop and planning to establish three new branches. The branches
should be allocated in such a way that the sum of distances from the customers to one of the three
branches should be kept minimum. You may assume your 15 customers are located in the
coordinates (x1, x2) as follows:
x1 4 5 8 5 5 2 5 3 6 7 10 8 10 9 10
x2 12 10 7 3 4 11 5 8 2 4 8 9 9 3 12
Calculate the locations of the three new branches.
Question 2
The file “Investment_portfolios.csv” contains 950 investment portfolios that include ten different
assets with the order as follows.
(i) Government Bond (GBD)
(ii) Listed Company Bond (LCB)
(iii) Gold (GLD)
(iv) Rare Earth Metal (REM)
(v) Stock Market 1 (SM1)
(vi) Stock Market 2 (SM2)
(vii) Oil & Gas (OGS)
(viii) Foreign Currency 1 (FC1)
(ix) Foreign Currency 2 (FC2)
(x) Mortgage (MOR)
(a) Illustrate how you could use the SimpleKMeans in Weka to identify the clusters if K = 3, 5, and 7.
The basic unit in the dataset is 1,000 US dollars.
(b) Explain what Silhouette coefficient (SC) is and justify what is the best choice of K in (a)
(c) List one of the major drawbacks in the SimpleKMeans learning algorithm.
(d) Explain what is meant by K-mean++ in cluster analysis.
(e) Identify the natural groups in the dataset using K-mean++ and find the centroids of the clusters if
K=3, 5, and 7.
*** End ***

学霸联盟

































学霸联盟