QUB | Archaeology and Palaeoecology | The 14Chrono Centre
e Correspondence analysis (off) [off] Enter `0' for on, or `1' for offThis type of multivariate analysis was developed to cope with enumerated data. It resembles principal components analysis, but is based on a different type of similarity matrix. Like other forms of ordination, results from correspondence analysis often plot along a marked curve, or horseshoe, when seen in two dimensions. Detrended Correspondence Analysis is correspondence analysis with a subsequent procedure to remove this tendency (i.e. detrend it). This part of the program is based on the FORTRAN program DECORANA, written by Hill 1979), with corrections from Oksanen & Minchin 1997).
Results are printed out to the data analysis file (see menu Ec), if one is available. The output includes eigenvalues, taxa loadings, and sample scores. These can, of course, be extracted and plotted in some appropriate manner.
It is often desirable to apply a threshold value (menu Mi), and recalculate the dataset for analysis to proportions of the sum of types included (menu Mj).
If 0 is selected, the following submenu appears:
a Detrended correspondence analysis (off) [off] b Downweighting of rare types (off) [off] c Rescaling of axes [off] d Number of times to rescale axes [4] e Rescaling threshold [0] f Number of segments [26] g Data transformation [0] h Modelling of ca (off) [off] 9 Leave this menu Q Return to main menu
Return to main menu.
Archaeology and Palaeoecology | 42 Fitzwilliam St | Belfast BT9 6AX | Northern Ireland | tel +44 28 90 97 5136