FREX scores for every word and topic

FREX balances word frequency and exclusivity (Bischof & Airoldi 2012; Roberts et al.). Unlike stm's labelTopics(), this returns the full numeric FREX matrix, not just the ranked words (addresses a long-standing stm request, bstewart/stm#265).

Usage

frex_scores(model, w = 0.5)

Arguments

model

A faSTM fit.

w

FREX frequency/exclusivity weight (0.5 = equal).

Value

A topics × vocabulary matrix of FREX scores (columns named by vocab).