Description
WordSegmenterModel (WSM) is based on maximum entropy probability model to detect word boundaries in Chinese text. Chinese text is written without white space between the words, and a computer-based application cannot know a priori which sequence of ideograms form a word. In many natural language processing tasks such as part-of-speech (POS) and named entity recognition (NER) require word segmentation as a initial step.
References:
- Xue, Nianwen. “Chinese word segmentation as character tagging.” International Journal of Computational Linguistics & Chinese Language Processing, Volume 8, Number 1, February 2003: Special Issue on Word Formation and Chinese Language Processing. 2003.).
How to use
Use as part of an nlp pipeline as a substitute of the Tokenizer stage.
document_assembler = DocumentAssembler() \
.setInputCol("text") \
.setOutputCol("document")
word_segmenter = WordSegmenterModel.pretrained('wordseg_kaist_ud', 'ko')\
.setInputCols("document")\
.setOutputCol("token")
pipeline = Pipeline(stages=[
document_assembler,
word_segmenter
])
model = pipeline.fit(spark.createDataFrame([[""]]).toDF("text"))
example = spark.createDataFrame([['비파를탄주하는그늙은명인의시는아름다운화음이었고완벽한음악으로순간적인조화를이룬세계의울림이었다.']], ["text"])
result = model.transform(example)
val document_assembler = DocumentAssembler()
.setInputCol("text")
.setOutputCol("document")
val word_segmenter = WordSegmenterModel.pretrained("wordseg_kaist_ud", "ko")
.setInputCols("document")
.setOutputCol("token")
val pipeline = new Pipeline().setStages(Array(document_assembler, word_segmenter))
val data = Seq("비파를탄주하는그늙은명인의시는아름다운화음이었고완벽한음악으로순간적인조화를이룬세계의울림이었다.").toDF("text")
val result = pipeline.fit(data).transform(data)
import nlu
text = ["""비파를탄주하는그늙은명인의시는아름다운화음이었고완벽한음악으로순간적인조화를이룬세계의울림이었다."""]
token_df = nlu.load('ko.segment_words').predict(text, output_level='token')
token_df
Results
+-------------------------------------------------------------------------------------------------+---------------------------------------------------------------------------------------------------------------------------------+
|text |result |
+-------------------------------------------------------------------------------------------------+---------------------------------------------------------------------------------------------------------------------------------+
|비파를탄주하는그늙은명인의시는아름다운화음이었고완벽한음악으로순간적인조화를이룬세계의울림이었다.|[비파를, 탄주하는, 그, 늙은, 명인의, 시는, 아름다운, 화음이었고, 완벽한, 음악으로, 순간적인, 조화를, 이룬, 세계의, 울림이었다, .]|
+-------------------------------------------------------------------------------------------------+---------------------------------------------------------------------------------------------------------------------------------+
Model Information
| Model Name: | wordseg_kaist_ud |
| Compatibility: | Spark NLP 2.7.0+ |
| Edition: | Official |
| Input Labels: | [document] |
| Output Labels: | [token] |
| Language: | ko |
Data Source
We trained the model using the Universal Dependenicies data set from Korea Advanced Institute of Science and Technology (KAIST-UD).
Reference:
Building Universal Dependency Treebanks in Korean, Jayeol Chun, Na-Rae Han, Jena D. Hwang, and Jinho D. Choi. In Proceedings of the 11th International Conference on Language Resources and Evaluation, LREC’18, Miyazaki, Japan, 2018.
Benchmarking
| Model | precision | recall | f1-score |
|---------------|--------------|--------------|--------------|
| KO_KAIST-UD | 0,6966 | 0,7779 | 0,7350 |