reset password
Author Message
xieguahu
Posts: 50
Posted 09:26 Jul 25, 2009 |

See the attachment

Attachments:
Last edited by xieguahu at 09:26 Jul 25, 2009.
nshatok
Posts: 19
Posted 14:23 Jul 25, 2009 |

Are we supposed to use Jaccard coefficient or Jaccard distance? The assignment says Jaccard coefficient , but the paper uses distance I think...

HelloWorld
Posts: 88
Posted 16:45 Jul 26, 2009 |
nshatok wrote:

Are we supposed to use Jaccard coefficient or Jaccard distance? The assignment says Jaccard coefficient , but the paper uses distance I think...

The question states:

9. Use the K-Mediods clustering algorithm with Jaccard Coefficient as the similarity measure ([SilvaLCT06] and lecture notes) to divide the transactions in Exercise 6 into two clusters.

But it seems like they're the same thing.. I tried using both the Jaccard Coefficient an the formula on the paper, they result in the same value..

I attached my solution, which is the same thing as GX's, but more detail..

Attachments:
nshatok
Posts: 19
Posted 17:15 Jul 26, 2009 |

In a paper they measure dissimilarity (variable d(a,b)), I understood,  but Jaccard coefficient measures similarity..

HelloWorld
Posts: 88
Posted 18:15 Jul 26, 2009 |
nshatok wrote:

In a paper they measure dissimilarity (variable d(a,b)), I understood,  but Jaccard coefficient measures similarity..

the dissimilarity is 1 - d

d is the same value as Jaccard coefficient

mbhatt
Posts: 14
Posted 23:27 Jul 26, 2009 |

I have used the following k mediod algo. as given in lec slides.

Following is the K mediod algorithm.

  1. Initialize: randomly select k of the n data points as the mediods
  2. Associate each data point to the closest medoid
  3. For each mediod m
    1. For each non-mediod data point o
      1. Swap m and o and compute the total cost of the configuration
  4. Select the configuration with the lowest cost.
  5. repeat steps 2 to 5 until there is no change in the medoid.              Answer is same but approach is different

 

Attachments:
Last edited by mbhatt at 23:59 Jul 26, 2009.