-n

Switch

-n N [N2 ...]

Description

The value(s) of n to use when extracting n-gram features with --add_ngrams/--add_ngrams_from_tokenized. Aliased as --set_n.

Argument and Default Value

One or more integers. Default: [1] (unigrams only).

Details

Verified: dlatkInterface.py -d dla_tutorial -t msgs -c user_id --add_ngrams -n 1 2 3 creates feat$1gram$..., feat$2gram$..., and feat$3gram$... tables, one per value of -n. Combine with --combine_feat_tables to also produce a single merged table across all of them (e.g. a 1to3gram table).

Other Switches

Required Switches:

• --add_ngrams or --add_ngrams_from_tokenized

Optional Switches:

• --combine_feat_tables

Example Commands

dlatkInterface.py -d dla_tutorial -t msgs -c user_id --add_ngrams -n 1 2 3 --combine_feat_tables 1to3gram