Predict method for k-gram frequency tables

Predictive text based on Stupid Back-off N-gram model.

# S3 method for sbo_kgram_freqs
predict(object, input, lambda = 0.4, ...)

Arguments

object	a sbo_kgram_freqs object.
input	a length one character vector, containing the input for next-word prediction.
lambda	a numeric vector of length one. The back-off penalization in Stupid Back-off algorithm.
...	further arguments passed to or from other methods.

Value

A tibble containing the next-word probabilities for all words in the dictionary.

Author

Valerio Gherardi

Examples

predict(twitter_freqs, "i love")
#> # A tibble: 1,001 x 2
#>    completion probability
#>    <chr>            <dbl>
#>  1 you             0.222 
#>  2 it              0.0574
#>  3 my              0.0549
#>  4 the             0.0524
#>  5 your            0.0312
#>  6 this            0.0312
#>  7 that            0.0287
#>  8 how             0.0249
#>  9 u               0.0175
#> 10 when            0.0150
#> # … with 991 more rows