Whether word-final /t, d/ is deleted (as in wes' side), 50 tokens from each of 40 speakers. Includes an interaction, a non-linear age effect, correlated predictors and two pure-noise predictors. For classification trees and random forests.
Format
A tibble with 2000 rows and 11 columns:
- speaker
S01toS40.- deleted
"retained"or"deleted".- following
Following segment:
"vowel","pause"or"consonant".- morph
"monomorpheme"(mist) or"past"(missed).- rate
Speech rate in syllables per second.
- log_freq
log10 word frequency, 1 to 6.
- age
Speaker age, 18 to 80.
- gender
"female"or"male"; no effect.- education
"school"or"university"; correlated with age but no effect of its own.- noise_cont, noise_bin
Random noise, continuous and binary; no effect.
Details
True model, in log-odds of deletion: -0.8 + s0 + 1.8 * consonant + 0.4 * pause - 1 * past + 0.9 * (rate - 5) * consonant + 1.2 * plogis((45 - age) / 5) + 0.25 * (log_freq - 3.5), with by-speaker
intercepts s0 of SD 0.5. Rate matters only before consonants; deletion
drops sharply around age 45.
Examples
if (requireNamespace("rpart", quietly = TRUE)) {
tree <- rpart::rpart(deleted ~ following + morph + rate + log_freq + age + gender +
education + noise_cont + noise_bin, data = td_deletion)
round(tree$variable.importance)
}
#> following age rate log_freq education morph noise_cont
#> 85 39 27 22 19 6 2
#> noise_bin
#> 1
if (requireNamespace("randomForest", quietly = TRUE)) {
set.seed(1)
rf <- randomForest::randomForest(deleted ~ following + morph + rate + log_freq + age +
gender + education + noise_cont + noise_bin,
data = td_deletion, importance = TRUE)
round(randomForest::importance(rf, type = 1), 1)
}
#> MeanDecreaseAccuracy
#> following 62.4
#> morph 22.6
#> rate 10.4
#> log_freq 10.1
#> age 28.9
#> gender 5.0
#> education 12.3
#> noise_cont 2.2
#> noise_bin -0.2
