Description
Classification of Self-Reported Intimate Partner Violence on Twitter. This model involves the detection the potential IPV victims on social media platforms (in English tweets).
Predicted Entities
intimate_partner_violence, non-intimate_partner_violence
Live Demo Open in Colab Copy S3 URI
How to use
document_assembler = DocumentAssembler() \
.setInputCol('text') \
.setOutputCol('document')
tokenizer = Tokenizer() \
.setInputCols(['document']) \
.setOutputCol('token')
sequenceClassifier = MedicalBertForSequenceClassification.pretrained("bert_sequence_classifier_self_reported_partner_violence_tweet", "en", "clinical/models")\
.setInputCols(["document",'token'])\
.setOutputCol("class")
pipeline = Pipeline(stages=[
document_assembler,
tokenizer,
sequenceClassifier
])
example = spark.createDataFrame(["I am fed up with this toxic relation.I hate my husband.",
"Can i say something real quick I ve never been one to publicly drag an ex partner and sometimes I regret that. I ve been reflecting on the harm, abuse and violence that was done to me and those bitches are truly lucky I chose peace amp therapy because they are trash forreal."], StringType()).toDF("text")
result = pipeline.fit(example).transform(example)
result.select("text", "class.result").show(truncate=False)
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_self_reported_partner_violence_tweet", "en", "clinical/models")
.setInputCols(Array("document","token"))
.setOutputCol("class")
val pipeline = new Pipeline().setStages(Array(document_assembler, tokenizer, sequenceClassifier))
# couple of simple examples
val example = Seq(Array("I am fed up with this toxic relation.I hate my husband.",
"Can i say something real quick I ve never been one to publicly drag an ex partner and sometimes I regret that. I ve been reflecting on the harm, abuse and violence that was done to me and those bitches are truly lucky I chose peace amp therapy because they are trash forreal.")).toDF("text")
val result = pipeline.fit(example).transform(example)
import nlu
nlu.load("en.classify.self_reported_partner_violence").predict("""Can i say something real quick I ve never been one to publicly drag an ex partner and sometimes I regret that. I ve been reflecting on the harm, abuse and violence that was done to me and those bitches are truly lucky I chose peace amp therapy because they are trash forreal.""")
Results
+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------------------------------+
|text |result |
+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------------------------------+
|I am fed up with this toxic relation.I hate my husband. |[non-intimate_partner_violence]|
|Can i say something real quick I ve never been one to publicly drag an ex partner and sometimes I regret that. I ve been reflecting on the harm, abuse and violence that was done to me and those bitches are truly lucky I chose peace amp therapy because they are trash forreal.|[intimate_partner_violence] |
+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------------------------------+
Model Information
| Model Name: | bert_sequence_classifier_self_reported_partner_violence_tweet |
| Compatibility: | Healthcare NLP 4.0.0+ |
| License: | Licensed |
| Edition: | Official |
| Input Labels: | [document, token] |
| Output Labels: | [class] |
| Language: | en |
| Size: | 406.5 MB |
| Case sensitive: | true |
| Max sentence length: | 128 |
References
Benchmarking
label precision recall f1-score support
intimate_partner_violence 0.96 0.97 0.97 630
non-intimate_partner_violence 0.75 0.69 0.72 78
accuracy - - 0.94 708
macro-avg 0.86 0.83 0.84 708
weighted-avg 0.94 0.94 0.94 708