I am trying to rank critical variables for 6 of the life cycle stages.
Hence I need to conduct Ranking of variables via covariance matrix before running Principal Component Analysis( dimension reduction)
In My data every variable is graded on a likert scale (1-5) and then its given either 1 or 0 weight for each life cycle stage.
(a) Plz advise how to first conduct weighted grading of each variable?
(b) Then how to finally conduct covariance based ranking?
Dear Asid
I am not sure whether this can help.
As you have a relevance ranking (1-5) and a max-min (1-0) criteria for each possible value between 1 and 5, perhaps you could apply Analytical Hierarchy Process (AHP) method in order to find eigenvalues for each variable.
Then you can proceed the PCA
The principal components of X are the eigenvectors of
CX = 1/n XXT (the resulting matriz multiplied by the Transposed Matrix) .
And the diagonal values of CY (considering X,Y as an orthogonal system) are the variances of X along the rows of the matrix that represents the orthogonal version of C.
Best regards,
If I understand you correctly then perhaps you
1. Can stretch your scale intuitively. You can have a 10 point scale with 1-5 for positive/yes/pass (1) and 6-10 for negative/no/fail (0).
2.Might also be interested in Cluster Analysis
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Applied Statistics
- Survey Design
- Questionnaire