Legal Research And Intellectual Property Document Classifier (EURLEX)

Description

European Union (EU) legislation is published in the EUR-Lex portal. All EU laws are annotated by the EU’s Publications Office with multiple concepts from the EuroVoc thesaurus, a multilingual thesaurus maintained by the Publications Office.

Given a document, the legclf_research_and_intellectual_property_bert model, it is a Bert Sentence Embeddings Document Classifier, classifies if the document belongs to the class Research_and_Intellectual_Property or not (Binary Classification) according to EuroVoc labels.

Predicted Entities

Research_and_Intellectual_Property, Other

Live Demo Open in Colab Download Copy S3 URI

How to use

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

embeddings = nlp.BertSentenceEmbeddings.pretrained("sent_bert_base_cased", "en")\
    .setInputCols("document")\
    .setOutputCol("sentence_embeddings")

doc_classifier = legal.ClassifierDLModel.pretrained("legclf_research_and_intellectual_property_bert", "en", "legal/models")\
    .setInputCols(["sentence_embeddings"])\
    .setOutputCol("category")

nlpPipeline = nlp.Pipeline(stages=[
    document_assembler, 
    embeddings,
    doc_classifier])

df = spark.createDataFrame([["YOUR TEXT HERE"]]).toDF("text")

model = nlpPipeline.fit(df)

result = model.transform(df)

Results

+-------+
|result|
+-------+
|[Research_and_Intellectual_Property]|
|[Other]|
|[Other]|
|[Research_and_Intellectual_Property]|

Model Information

References

Train dataset available here

Benchmarking

                             label precision recall  f1-score  support
                             Other      0.80   0.82      0.81      110
Research_and_Intellectual_Property      0.85   0.83      0.84      133
                          accuracy         -      -      0.83      243
                         macro-avg      0.83   0.83      0.83      243
                      weighted-avg      0.83   0.83      0.83      243