data normalization a priori to analysis can be quite difficult. We suggest to employ analysis tools, which are Independent of the Chosen normalization, which means they don't need any data normalization at all. Have a look at our zero-sum regression publications.
Dear colleagues,
We use the Johnson translation functions for data normalization. Unlike well-known the Box-Cox transformation, these functions can be chosen on the bases to the values of skewness square and kurtosis before normalization.
In addition see my answers in the link below.
https://www.researchgate.net/post/Log_transformation_to_construct_non-normal_data_as_normal-How_far_it_is_justified_for_statistical_analysis
Regards, Sergiy
I think the previous answer is good if you want to do this as it uses more than the first two moments for the normalization conversion. However, I suggest you be very careful before preforming a normalization process before analyzing data as you will also be modifying your results. If your data are not approximately normal and the sample size is too small to overcome the small size, your inferences will be for normal data that do not exist. Thus, your results will be suspect.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression