Part of Speech for Icelandic

Description

A Part of Speech classifier predicts a grammatical label for every token in the input text. Implemented with an averaged perceptron architecture.

Predicted Entities

  • ADJ
  • ADP
  • ADV
  • AUX
  • CCONJ
  • DET
  • NOUN
  • NUM
  • PART
  • PRON
  • PROPN
  • PUNCT
  • VERB
  • X

Live Demo Open in Colab Download

How to use

document_assembler = DocumentAssembler()\
  .setInputCol("text")\
  .setOutputCol("document")

sentence_detector = SentenceDetector()\
  .setInputCols(["document"])\
  .setOutputCol("sentence")

pos = PerceptronModel.pretrained("pos_icepahc", "is")\
  .setInputCols(["document", "token"])\
  .setOutputCol("pos")

pipeline = Pipeline(stages=[
  document_assembler,
  sentence_detector,
  posTagger
])

example = spark.createDataFrame(pd.DataFrame({'text': ["Númerið blikkaði á skjánum eins og einmana vekjaraklukka um nótt á níundu hæð í gamalli blokk í austurbæ Reykjavíkur ."]}))
result = pipeline.fit(example).transform(example)
val document_assembler = DocumentAssembler()
        .setInputCol("text")
        .setOutputCol("document")

val sentence_detector = SentenceDetector()
        .setInputCols(["document"])
	.setOutputCol("sentence")

val pos = PerceptronModel.pretrained("pos_icepahc", "is")
        .setInputCols(Array("document", "token"))
        .setOutputCol("pos")

val pipeline = new Pipeline().setStages(Array(document_assembler, sentence_detector, pos))

val result = pipeline.fit(Seq.empty["Númerið blikkaði á skjánum eins og einmana vekjaraklukka um nótt á níundu hæð í gamalli blokk í austurbæ Reykjavíkur ."].toDS.toDF("text")).transform(data)
import nlu

text = [""Númerið blikkaði á skjánum eins og einmana vekjaraklukka um nótt á níundu hæð í gamalli blokk í austurbæ Reykjavíkur .""]
token_df = nlu.load('is.pos.icepahc').predict(text)
token_df

Results

+----------------------------------------------------------------------------------------------------------------------+-----------------------------------------------------------------------------------------------------------------+
|text                                                                                                                  |result                                                                                                           |
+----------------------------------------------------------------------------------------------------------------------+-----------------------------------------------------------------------------------------------------------------+
|Númerið blikkaði á skjánum eins og einmana vekjaraklukka um nótt á níundu hæð í gamalli blokk í austurbæ Reykjavíkur .|[NOUN, VERB, ADP, NOUN, ADV, ADP, ADJ, NOUN, ADP, NOUN, ADP, ADJ, NOUN, ADP, ADJ, NOUN, ADP, PROPN, PROPN, PUNCT]|
+----------------------------------------------------------------------------------------------------------------------+-----------------------------------------------------------------------------------------------------------------+

Model Information

Model Name: pos_icepahc
Compatibility: Spark NLP 2.7.5+
License: Open Source
Edition: Official
Input Labels: [sentence, token]
Output Labels: [pos]
Language: is

Data Source

The model was trained on the Universal Dependencies data set.

Benchmarking

|              | precision | recall | f1-score | support |
|--------------|-----------|--------|----------|---------|
| ADJ          | 0.81      | 0.74   | 0.78     | 5906    |
| ADP          | 0.95      | 0.96   | 0.96     | 15548   |
| ADV          | 0.90      | 0.90   | 0.90     | 10631   |
| AUX          | 0.92      | 0.93   | 0.92     | 7416    |
| CCONJ        | 0.96      | 0.97   | 0.96     | 8437    |
| DET          | 0.89      | 0.87   | 0.88     | 7476    |
| INTJ         | 0.95      | 0.77   | 0.85     | 131     |
| NOUN         | 0.90      | 0.92   | 0.91     | 20726   |
| NUM          | 0.75      | 0.83   | 0.79     | 655     |
| PART         | 0.96      | 0.96   | 0.96     | 1703    |
| PRON         | 0.94      | 0.96   | 0.95     | 16852   |
| PROPN        | 0.89      | 0.89   | 0.89     | 4444    |
| PUNCT        | 0.98      | 0.98   | 0.98     | 16434   |
| SCONJ        | 0.94      | 0.94   | 0.94     | 5663    |
| VERB         | 0.92      | 0.90   | 0.91     | 17329   |
| X            | 0.60      | 0.30   | 0.40     | 346     |
| accuracy     |           |        | 0.92     | 139697  |
| macro avg    | 0.89      | 0.86   | 0.87     | 139697  |
| weighted avg | 0.92      | 0.92   | 0.92     | 139697  |