Alternatively, you may prefer to use command lines to compute correspondence analysis. So you can load your own data and use the following commands:
>> addpath lib;
>> output = canalysis(N,k,labels_x,labels_y);
>> rotation = ComputeRotation(output, k, coordinates, method, weights);
>> PrintDescriptives(output,detailed);
>> PrintCoordinates(output,k,coordinates,detailed);
>> PrintRotation(rotation,method, weights, output.labels, detailed);
>> map(1,2,0);where
(1) 'N' is a contingence table,
(2) 'k' is the number of dimensions to be retained.
(3) 'labels_x' and 'labels_y' store the corresponding labels,
(4) 'coordinates' is the coordinates used in the model. The admissible values are:
| 1 | Symmetrical coordinates (biplot model) | |
| 2 | Row principal coordinates, and column standard coordinates | |
| 3 | Column principal coordinates, and row standard coordinates | |
| 4 | Row and column principal coordinates (French symmetrical model) |
(5) 'method' is the rotation method to be computed. The admissible values are:
| 1 | Orthogonal rotation of symmetrical coordinates | |
| 2 | Oblique rotation of symmetrical coordinates | |
| 3 | Orthogonal rotation of principal coordinates | |
| 4 | Orthogonal rotation of standard coordinates | |
| 5 | Oblique rotation of principal coordinates |
(6) 'weights' is the weight scheme to be used in the rotation. The admissible values are:
| 1 | Row-wise normalization of coordinates | |
| 2 | Rescaling coordinates using masses | |
| 3 | None. No weights scheme is used. |
(7) 'detailed' controls the level of detail of the report. The admissible values are:
| 1 | Prints all the details | |
| 0 | Prints the basic information |
If you want to include some supplementary rows or columns, then you should replace this line:
>> output = canalysis(N,k,labels_x,labels_y);for this one:
>> output = canalysis(N,k,labels_x,labels_y, Sup, labels_sup);where
(8) 'Sup' is a structure that contains matrices 'r' and/or 'c' that are the contingence tables which contain the supplementary rows and columns.
(9) 'labels' is a structure that stores the labels for the supplementary rows and columns in strings 'r' and 'c' (for rows and columns, respectively).
Finally, function 'canalysis' can be also executed as,
>> output = canalysis(N);In this case, the number of dimensions to retain is set to 2, and default labels for rows and columns are used.