Re: 2 Questions: Manova and Selecting features

Thank you so much for your reply ... :)

Antonio wrote:


Hi,

Also you should verify the following MANOVA files

Computes a Multivariate Analysis of Variance for equal or unequal
sample sizes.
<http://www.mathworks.com/matlabcentral/fileexchange/loadFile.do?objectId=3863>

Computes a two-way Multivariate Analysis of Variance for equal or
unequal sample sizes.
<http://www.mathworks.com/matlabcentral/fileexchange/loadFile.do?objectId=4073>

Statistical power of a performed (a posteriori) single-factor
MANOVA.
<http://www.mathworks.com/matlabcentral/fileexchange/loadFile.do?objectId=3864>

Calculation of the a priori statistical power of a one-way MANOVA
test.
<http://www.mathworks.com/matlabcentral/fileexchange/loadFile.do?objectId=3880>

--Regarding to your question on the best way to choose the most
important attributes ? (Except PCA !!).

Well, one other option is to use the Cluster Analysis. Clustering
and
partitioning methods are used to group cases on the basis of their
similarity over a range of variables or variables over a range of
cases. The main examples of these techniques come under the general
heading of cluster analysis. Many clustering algorithms are
available; they differ with respect to the method used to measure
similarities (or dissimilarities) and the points between which
distances are measured. Thus, although clustering algorithms are
objective, there is scope for subjectivity in the selection of an
algorithm. The most common clustering algorithms are polythetic
agglomerative, i.e. a series of increasingly larger clusters are
formed by the fusion of smaller clusters on the basis of more than
one variable. A problem with the hierarchical approach is that they
are computer-intensive and large data sets may be difficult to
analyzing. A less computer intensive approach is the
nonhierarchical
k means or iterative relocation algorithm. Each case is initially
placed in one of k clusters, cases are then moved between clusters
if
it minimizes the differences between cases within a cluster.

Cheers,
Antonio
.

Relevant Pages

  • Re: 2 Questions: Manova and Selecting features
    ... Computes a Multivariate Analysis of Variance for equal or unequal ... Statistical power of a performed single-factor MANOVA. ... Many clustering algorithms are ... Thus, although clustering algorithms are ...
    (comp.soft-sys.matlab)
  • Re: Feature selection and K-means clustering
    ... Function 'sequentialfs' can be applied for both supervised learning ... unsupervised algorithms (such as clustering algorithms). ... Selecting features for clustering is not easy in general. ...
    (comp.soft-sys.matlab)
  • Re: datasets to test clustering algorithm
    ... > use in the comparison of a few clustering algorithms? ... see if you end up clustering books by the same author together, ... And, if your clustering algorithms get better results than Hoover got, ...
    (sci.stat.math)
  • Re: assumption of Classification
    ... Most clustering algorithms (k-means, ... > clustering algorithms which assumes that each cluster is multivariate ... A classification tree that tries to break at every value ... normality is not the only assumption to be checked. ...
    (sci.stat.edu)