Use of correspondence analysis partial least squares on linear and unimodal data.

Correspondence analysis partial least squares (CA-PLS) has been compared with PLS concerning classification and prediction of unimodal growth temperature data and an example using infrared (IR) spectroscopy for predicting amounts of chemicals in mixtures. CA-PLS was very effective for ordinating the unimodal temperature data and the results indicated that CA-PLS is effective in treating the arch effect, thus avoiding the detrending procedure often used on ecological data sets, at least when one basic underlying gradient is present. PLS and PCR gave poor results, as the ordinations had a horseshoe form that could only be seen in two-dimensional plots, and also less effective predictions. PLS was the best method in the linear case treated, with fewer components and a better prediction than CA-PLS.