Docstoc

mrSess_25

Document Sample
mrSess_25 Powered By Docstoc
					WELCOME
  FRIENDS
Multivariate measures of
       association

• Factor Analysis
• Cluster Analysis




                     Cluster Analysis
       Factor analysis
• The term factor analysis was first introduced
  by Thurston in 1931.
• The main applications of this technique are:
  1- To reduce the number of variables to a
     manageable number.
  2- To detect the structure in the relationship
     between variables.
         Factor analysis
• METHOD: It can be done in many ways. The
    approach used mostly principle component analysis. In
    this the given set of variables is reduced to another set
    of composite variable.
•   These variables are not correlated to one another. These
    new variables known as factors account for the variance
    of entire data.
•   The best combination of variables forms the first factor
    & accounts for most of variance.
•   The next factor is chosen such that it is the best
    combination that accounts for variance unexplained by
    the first factor.
•   This process is repeated until all the variance is
    accounted for, but researchers stops after most of the
    variances accounted for.
 Cluster analysis
• Cluster analysis was developed to cater to
  getting the answer of- How to organize
  observed data into meaningful structure.
• Unlike Discriminant analysis begins with it
  doesn’t have prior hypothesis.
• Cluster analysis starts with an
  undifferentiated group of people, events &
  objects & recognizes them into homogenous
  group.
    Cluster analysis
• Method: Once the sample is selected the following
    steps are through gone through to arrive @ cluster.
•   Variables on which objects, events or people are to be
    measured should be defined.
•   Similarities among the elements should be calculated.
•   Mutually exclusive groups (cluster) should be made.
•   These clusters should be compared with each other &
    they could be validated.
•   Different cluster methods result in different results so the
    data must be understood well to differentiate
    between the real & imposed cluster.
  THANKS
• BY:
    SANDEEP SAXENA
    FACULTY MEMBER MARKETING
    INC- HQ, HYDERABAD

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:0
posted:9/15/2012
language:
pages:7