+ Reply to Thread
Results 1 to 5 of 5

Thread: Principle Components Analysis and Cluster Analysis

  1. #1
    Points: 755, Level: 14
    Level completed: 55%, Points required for next Level: 45

    Posts
    9
    Thanks
    4
    Thanked 0 Times in 0 Posts

    Principle Components Analysis and Cluster Analysis




    Hi,

    I was wondering if you could all hep me decode the difference between both PCA and cluster analysis. As far as I can tell, PCA aims toreduc/cluster the variables, but cluster analysis attempts to cluster the results?

    Thanks

    QOS

  2. #2
    TS Contributor
    Points: 4,687, Level: 43
    Level completed: 69%, Points required for next Level: 63
    Miner's Avatar
    Location
    Greater Milwaukee area
    Posts
    548
    Thanks
    23
    Thanked 186 Times in 165 Posts

    Re: Principle Components Analysis and Cluster Analysis

    Principle Components Analysis (PCA) and Factor Analysis are data reduction techniques. The objective to to reduce large numbers of variables to a smaller number of "factors", or underlying constructs that explain the variation. For example, you have data on gripping strength of each hand, lower and upper arm strength. You have the same for different measures of leg strength. PCA or Factor Analysis might reduce this to two constructs called upper body strength and lower body strength. Both methods focus on the Independent Variables.

    Cluster Analysis may be run on raw data, scores from Factor Analysis, or on variables. For the first two, CA will find clusters of entities (think market segments) that behave in a similar fashion. For the latter, CA will find variables that measure similar responses. This is particularly useful to identify survey questions that are asking the same thing.

  3. The Following User Says Thank You to Miner For This Useful Post:

    QueenOfStats (05-07-2013)

  4. #3
    Points: 1,249, Level: 19
    Level completed: 49%, Points required for next Level: 51

    Location
    Bangalore
    Posts
    135
    Thanks
    3
    Thanked 30 Times in 29 Posts

    Re: Principle Components Analysis and Cluster Analysis

    Another way of understanding is that PCA factors the variables whereas Clustering works on rows of data.

  5. The Following User Says Thank You to chetan.apa For This Useful Post:

    QueenOfStats (05-07-2013)

  6. #4
    Points: 755, Level: 14
    Level completed: 55%, Points required for next Level: 45

    Posts
    9
    Thanks
    4
    Thanked 0 Times in 0 Posts

    Re: Principle Components Analysis and Cluster Analysis

    Perfect, thanks for the help!
    So they're quite similar then, I guess, but can both approach things with a different angle. Cluster analysis isn't really reducing data, but just finding a way to capture the scores. But PCA is reducing the variables rather than the outcomes, right? This **** is intense

  7. #5
    TS Contributor
    Points: 4,687, Level: 43
    Level completed: 69%, Points required for next Level: 63
    Miner's Avatar
    Location
    Greater Milwaukee area
    Posts
    548
    Thanks
    23
    Thanked 186 Times in 165 Posts

    Re: Principle Components Analysis and Cluster Analysis


    Quote Originally Posted by QueenOfStats View Post
    But PCA is reducing the variables rather than the outcomes, right?
    That is correct.

  8. The Following User Says Thank You to Miner For This Useful Post:

    QueenOfStats (05-07-2013)

+ Reply to Thread

           




Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats