-
Notifications
You must be signed in to change notification settings - Fork 6
Home
Dima Kudosh edited this page Feb 3, 2016
·
2 revisions
Firstly import word2vec.
extern crate word2vec;
Then you can create WordVector object using load_from_binary method. It has only one parameter - file_name and it return Result<WordVector, Word2VecError>
.
let wordvectors = word2vec::wordvectors::WordVector::load_from_binary("vectors.bin");
WordVector object has 3 public methods:
-
get_vector(&self, word: &str) -> Option<&Vec<f32>>
Return vector for specified word or None -
cosine(&self, word: &str, n: usize) -> Option<Vec<(String, f32)>>
Return n tuples with word and probability or None -
analogy(&self, pos: Vec<&str>, neg: Vec<&str>, n: usize) -> Option<Vec<(String, f32)>>
Return n tuples with word and probability or None
Also you can load word clusters from file. For this you must create WordClusters object using load_from_file method:
let clusters = word2vec::wordclusters::WordClusters::load_from_file("classes.txt");
It's takes only one parameter (filename: &str) and return Result<WordClusters, Word2VecError>
WordClusters has 2 public method:
get_words_on_cluster(&self, index: i32) -> Option<&Vec<String>>
get_cluster(&self, word: &str) -> Option<&i32>