--p_correction

Switch

--p_correction

Description

Specifies a p-value correction method (for multiple comparisons) in correlation other than Bonferroni (which is turned off with --no_bonferroni)

Argument and Default Value

Argument: method to use

Details

Possible values include: simes, holm, hochberg, hommel, bonferroni, BH, BY, fdr, none

simes is built into DLATK directly; anything else calls R's stats module (via rpy2), specifically the p_adjust command — so any value other than simes requires rpy2 to be installed. Default correction method (if this flag is not given) is Benjamini-Hochberg (BH). To turn off correction entirely use --no_correction, not --no_bonferroni (which is currently broken — see that page).

Other Switches

Required Switches:

• --correlate

Example Commands

Using the packaged dla_tutorial dataset (see Differential Language Analysis (DLA) Tutorial):

dlatkInterface.py -d dla_tutorial -t msgs -c user_id -f 'feat$1gram$msgs$user_id$16to16' \
--outcome_table blog_outcomes --outcomes age --correlate --p_correction simes