|
•
|
r and c are row and column sums of P
|
|
•
|
C is the Burt table.
|
|
•
|
Q is the number of categorical variables
|
|
•
|
n is the number of observations
|
The usual principal inertias of a Burt table constructed from m categorical variables in MCA are the eigenvalues uk from 
. These inertias provide a pessimistic indication of fit. Benzécri (1979) proposed the following inertia adjustment; it is also described by Greenacre (1984, p. 145):
. These inertias provide a pessimistic indication of fit. Benzécri (1979) proposed the following inertia adjustment; it is also described by Greenacre (1984, p. 145):
for
.
for all inertias greater than 
, where 
is the sum of squared inertias and nc is the total number of categories across the m variables.
, where is the sum of squared inertias and nc is the total number of categories across the m variables.Quality is the sum of the squared cosines in the two dimensions. It is also equal to the ratio of the sum of inertias at two dimensions to the sum of the inertias in all dimensions. Quality indicates how well the point is represented in the two-dimension space.
Mass is the proportion of row or column total frequency to the total frequency.
Inertia is analogous to variance in principal component analysis. The overall inertia is the total Pearson Chi-square for a two-way frequency table divided by the sum of all observations in the table.
Relative inertia is the proportion of the contribution of the point to the overall inertia. In the summary statistics table, the relative inertia is listed in the column labeled Inertia.
