HIGHLIGHTS
SUMMARY
For each parameter, four types of data conditioning were compared, namely, the raw data, a logarithmic transformation, a square root transformation, and a transformation aiming to reinforce the role of the chemical profile independently of the dissolved load. Feng et_al mentioned that if the original data follow a log-normal or approximately normal distribution, then the log-transformed data follow a normal or approximately normal distribution and the log transformation removes or reduces the skewness. The main difference between PCA on raw data and transformed data, and in particular on log-transformed data, is . . .
If you want to have access to all the content you need to log in!
Thanks :)
If you don't have an account, you can create one here.