Question

Include one or two methods that could be listed on both tables as suitable for some general clustering problems as well as some rearrangement clustering problems.

a. Which method may be suitable for both general and rearrangement clustering problems?

b. What are the pros and cons of using this method for rearrangement clustering?

c. What are the pros and cons of using this method for general clustering?

Answer 1

cluster means set of technique used to partition a set of objectives into relatively homogenous subsets based on similarity

a. they are different type of clustering methods are used:

• hierarchial clustering
• density- based clustering
• model based clustering

b. advantage

• relatively simple to implement
• scales to large data sets
• easily adapt to new examples

disadvantage

• it is difficult to estimate the appropriate number of cluster in advance.
• being dependent on initial values
• clustering data of varying sizes and density
• scaling with number of dimensions

c. advantage

• no need define number of cluster in advance
• help to calculate whole hierarchy of clusters
• uses of dendogram for graphical representation
• it discover cluster of arbitary shapes.

disadvantage

• inability to make correction once the splitting/ merging decision is made
• lack of interpretability regarding the cluster descriptions
• high sensitive to setting of input parameters
• poor cluster descriptions