Plot the LDA topic count sweep

The three selection metrics and the combined score against k. Look for the k where the combined score peaks, then sanity check that Cao Juan has not already started climbing, which is the signal that topics are duplicating.

Usage

# S3 method for class 'LdaKSweepResult'
plot(x, ...)

Arguments

x

LdaKSweepResult object.

...

Ignored.

Value

A ggplot2 object with one panel per metric.

Examples

# the three selection metrics and the combined score against k
set.seed(42L)
corpus <- matrix(rbinom(200L * 40L, 1L, 0.05), nrow = 200L, ncol = 40L)
corpus[1:100, 1:10] <- rbinom(1000L, 1L, 0.6)
corpus[101:200, 11:20] <- rbinom(1000L, 1L, 0.6)
colnames(corpus) <- sprintf("term_%02d", 1:40)
sweep_res <- lda_k_sweep(corpus > 0, k_range = 5:7, .verbose = FALSE)
plot(sweep_res)