--set_p_occ

Switch

--set_p_occ <pocc_thresh>

Description

Meant to be used with --feat_occ_filter. See there for more info.

Argument and Default Value

pocc_thresh — threshold for the percentage of groups (that are above --group_freq_thresh wordcount) that must use a given feature for it to be kept. Default: 0.01 (dlac.DEF_P_OCC).

Other Switches

Required Switches:

• --feat_occ_filter

Example Commands

Using the packaged dla_tutorial dataset (see Differential Language Analysis (DLA) Tutorial): keeps only features used by at least 5% of groups.

dlatkInterface.py -d dla_tutorial -t msgs -c user_id -f 'feat$1to3gram$msgs$user_id' \
--feat_occ_filter --set_p_occ 0.05 --group_freq_thresh 500