Description
This model is a BioBERT based classifier that can classify premise about health mandates related to Covid-19 from tweets. This model is intended for direct use as a classification model and the target classes are: Has no premise (no argument), Has premise (argument).
Predicted Entities
Has premise (argument), Has no premise (no argument)
How to use
document_assembler = DocumentAssembler() \
.setInputCol("text") \
.setOutputCol("document")
tokenizer = Tokenizer() \
.setInputCols(["document"]) \
.setOutputCol("token")
sequence_classifier = MedicalBertForSequenceClassification.pretrained("bert_sequence_classifier_health_mandates_premise_tweet_onnx", "en", "clinical/models")\
.setInputCols(["document", "token"])\
.setOutputCol("class")
pipeline = Pipeline(stages=[
document_assembler,
tokenizer,
sequence_classifier
])
data = spark.createDataFrame(["""It's too dangerous to hold the RNC, but let's send students and teachers back to school.""",
"""So is the flu and pneumonia what are their s stop the Media Manipulation covid has treatments You're Speaker Pelosi nephew so stop the agenda LIES.""",
"""Just a quick update to my U.S. followers, I'll be making a stop in all 50 states this spring! No tickets needed, just don't wash your hands, cough on each other.""",
"""Go to a restaurant no mask Do a food shop wear a mask INCONSISTENT No Masks No Masks.""",
"""But if schools close who is gonna occupy those graves Cause politiciansprotected smokers protected drunkardsprotected school kids amp teachers""",
"""New title Maskhole I think I'm going to use this very soon coronavirus."""
], StringType()).toDF("text")
model = pipeline.fit(data)
result = model.transform(data)
document_assembler = nlp.DocumentAssembler() \
.setInputCol("text") \
.setOutputCol("document")
tokenizer = nlp.Tokenizer() \
.setInputCols(["document"]) \
.setOutputCol("token")
sequenceClassifier = medical.BertForSequenceClassification.pretrained("bert_sequence_classifier_health_mandates_premise_tweet_onnx", "en", "clinical/models")\
.setInputCols(["document","token"])\
.setOutputCol("classes")
pipeline = nlp.Pipeline(stages=[
document_assembler,
tokenizer,
sequenceClassifier
])
data = spark.createDataFrame(["""It's too dangerous to hold the RNC, but let's send students and teachers back to school.""",
"""So is the flu and pneumonia what are their s stop the Media Manipulation covid has treatments You're Speaker Pelosi nephew so stop the agenda LIES.""",
"""Just a quick update to my U.S. followers, I'll be making a stop in all 50 states this spring! No tickets needed, just don't wash your hands, cough on each other.""",
"""Go to a restaurant no mask Do a food shop wear a mask INCONSISTENT No Masks No Masks.""",
"""But if schools close who is gonna occupy those graves Cause politiciansprotected smokers protected drunkardsprotected school kids amp teachers""",
"""New title Maskhole I think I'm going to use this very soon coronavirus."""
], StringType()).toDF("text")
model = pipeline.fit(data)
result = model.transform(data)
val document_assembler = new DocumentAssembler()
.setInputCol("text")
.setOutputCol("document")
val tokenizer = new Tokenizer()
.setInputCols(Array("document"))
.setOutputCol("token")
val sequenceClassifier = MedicalBertForSequenceClassification.pretrained("bert_sequence_classifier_health_mandates_premise_tweet_onnx", "en", "clinical/models")
.setInputCols(Array("document","token"))
.setOutputCol("class")
val pipeline = new Pipeline().setStages(Array(document_assembler, tokenizer, sequenceClassifier))
val data = Seq(
"It's too dangerous to hold the RNC, but let's send students and teachers back to school.",
"So is the flu and pneumonia what are their s stop the Media Manipulation covid has treatments You're Speaker Pelosi nephew so stop the agenda LIES.",
"Just a quick update to my U.S. followers, I'll be making a stop in all 50 states this spring! No tickets needed, just don't wash your hands, cough on each other.",
"Go to a restaurant no mask Do a food shop wear a mask INCONSISTENT No Masks No Masks.",
"But if schools close who is gonna occupy those graves Cause politiciansprotected smokers protected drunkardsprotected school kids amp teachers",
"New title Maskhole I think I'm going to use this very soon coronavirus."
).toDF("text")
val model = pipeline.fit(data)
val result = model.transform(data)
Results
+------------------------------------------------------------------------------------------------------------------------------------------------------------------+------------------------------+
|text |result |
+------------------------------------------------------------------------------------------------------------------------------------------------------------------+------------------------------+
|It's too dangerous to hold the RNC, but let's send students and teachers back to school. |[Has premise (argument)] |
|So is the flu and pneumonia what are their s stop the Media Manipulation covid has treatments You're Speaker Pelosi nephew so stop the agenda LIES. |[Has premise (argument)] |
|Just a quick update to my U.S. followers, I'll be making a stop in all 50 states this spring! No tickets needed, just don't wash your hands, cough on each other.|[Has no premise (no argument)]|
|Go to a restaurant no mask Do a food shop wear a mask INCONSISTENT No Masks No Masks. |[Has no premise (no argument)]|
|But if schools close who is gonna occupy those graves Cause politiciansprotected smokers protected drunkardsprotected school kids amp teachers |[Has premise (argument)] |
|New title Maskhole I think I'm going to use this very soon coronavirus. |[Has no premise (no argument)]|
+------------------------------------------------------------------------------------------------------------------------------------------------------------------+------------------------------+
Model Information
| Model Name: | bert_sequence_classifier_health_mandates_premise_tweet_onnx |
| Compatibility: | Healthcare NLP 6.1.1+ |
| License: | Licensed |
| Edition: | Official |
| Input Labels: | [document, token] |
| Output Labels: | [label] |
| Language: | en |
| Size: | 437.7 MB |
| Case sensitive: | true |