Fake News Classifier - Spark NLP 2.7.1+

Description

Determine if news articles are Real or Fake.

Predicted Entities

REAL, FAKE

Live Demo Open in Colab Download

How to use

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

use = UniversalSentenceEncoder.pretrained('tfhub_use', lang="en") \
  .setInputCols(["document"])\
  .setOutputCol("sentence_embeddings")

document_classifier = ClassifierDLModel.pretrained('classifierdl_use_fakenews', 'en') \
  .setInputCols(["document", "sentence_embeddings"]) \
  .setOutputCol("class")

nlpPipeline = Pipeline(stages=[document_assembler, use, document_classifier])
light_pipeline = LightPipeline(nlp_pipeline.fit(spark.createDataFrame([['']]).toDF("text")))

annotations = light_pipeline.fullAnnotate('Donald Trump a KGB Spy? 11/02/2016 In today’s video, Christopher Greene of AMTV reports Hillary Clinton')

Results

+--------------------------------------------------------------------------------------------------------+------------+
|document                                                                                                |class       |
+--------------------------------------------------------------------------------------------------------+------------+
|Donald Trump a KGB Spy? 11/02/2016 In today’s video, Christopher Greene of AMTV reports Hillary Clinton | FAKE       |
+--------------------------------------------------------------------------------------------------------+------------+

Model Information

Model Name: classifierdl_use_fakenews
Compatibility: Spark NLP 2.7.1+
License: Open Source
Edition: Official
Input Labels: [sentence_embeddings]
Output Labels: [class]
Language: en
Dependencies: tfhub_use

Data Source

This model is trained on the fake new classification challenge. https://raw.githubusercontent.com/joolsa/fake_real_news_dataset/master/fake_or_real_news.csv.zip

Benchmarking

              precision    recall  f1-score   support

        FAKE       0.86      0.89      0.88       626
        REAL       0.89      0.86      0.87       634

    accuracy                           0.87      1260
   macro avg       0.88      0.87      0.87      1260
weighted avg       0.88      0.87      0.87      1260