Description
This Financial NER model can extract up to 20 quantifiable entities, including KPI, from the Responsibility and ESG Reports of companies. It has been trained with the SOTA approach.
Predicted Entities
AGE, AMOUNT, COUNTABLE_ITEM, DATE_PERIOD, ECONOMIC_ACTION, ECONOMIC_KPI, ENVIRONMENTAL_ACTION, ENVIRONMENTAL_KPI, ENVIRONMENTAL_UNIT, ESG_ROLE, FACILITY_PLACE, ISO, PERCENTAGE, PROFESSIONAL_GROUP, RELATIVE_METRIC, SOCIAL_ACTION, SOCIAL_KPI, TARGET_GROUP, TARGET_GROUP_BUSINESS, WASTE
How to use
document_assembler = nlp.DocumentAssembler()\
.setInputCol("text")\
.setOutputCol("document")\
sentence_detector = nlp.SentenceDetector()\
.setInputCols(["document"])\
.setOutputCol("sentence")\
tokenizer = nlp.Tokenizer() \
.setInputCols(["sentence"]) \
.setOutputCol("token")\
.setContextChars(['.', ',', ';', ':', '!', '?', '*', '-', '(', ')', '"', "'", '%', '&'])
ner_model = finance.BertForTokenClassification.pretrained("finner_responsibility_reports", "en", "finance/models")\
.setInputCols(["sentence", "token"])\
.setOutputCol("ner")\
.setCaseSensitive(True)\
.setMaxSentenceLength(512)
ner_converter = nlp.NerConverter()\
.setInputCols(["sentence", "token", "ner"])\
.setOutputCol("ner_chunk")
pipeline = nlp.Pipeline(stages=[
document_assembler,
sentence_detector,
tokenizer,
ner_model,
ner_converter
])
empty_data = spark.createDataFrame([[""]]).toDF("text")
model = pipeline.fit(empty_data)
text = """The company has reduced its direct GHG emissions from 12,135 million tonnes of CO2e in 2017 to 4 million tonnes of CO2e in 2021. The indirect GHG emissions (scope 2) are mainly from imported energy, including electricity, heat, steam, and cooling, and the company has reduced its scope 2 emissions from 3 million tonnes of CO2e in 2017-2018 to 4 million tonnes of CO2e in 2020-2021. The scope 3 emissions are mainly from the use of sold products, and the emissions have increased from 377 million tonnes of CO2e in 2017 to 408 million tonnes of CO2e in 2021."""
data = spark.createDataFrame([[text]]).toDF("text")
result = model.transform(data)
result.select(F.explode(F.arrays_zip('ner_chunk.result', 'ner_chunk.metadata')).alias("cols")) \
.select(F.expr("cols['0']").alias("chunk"),
F.expr("cols['1']['entity']").alias("label")).show(50, truncate = False)
Results
+----------------------+------------------+
|chunk |label |
+----------------------+------------------+
|direct GHG emissions |ENVIRONMENTAL_KPI |
|12,135 million |AMOUNT |
|tonnes of CO2e |ENVIRONMENTAL_UNIT|
|2017 |DATE_PERIOD |
|4 million |AMOUNT |
|tonnes of CO2e |ENVIRONMENTAL_UNIT|
|2021 |DATE_PERIOD |
|indirect GHG emissions|ENVIRONMENTAL_KPI |
|scope 2 |ENVIRONMENTAL_KPI |
|imported energy |ENVIRONMENTAL_KPI |
|electricity |ENVIRONMENTAL_KPI |
|heat |ENVIRONMENTAL_KPI |
|steam |ENVIRONMENTAL_KPI |
|cooling |ENVIRONMENTAL_KPI |
|scope 2 emissions |ENVIRONMENTAL_KPI |
|3 million |AMOUNT |
|tonnes of CO2e |ENVIRONMENTAL_UNIT|
|2017-2018 |DATE_PERIOD |
|4 million |AMOUNT |
|tonnes of CO2e |ENVIRONMENTAL_UNIT|
|2020-2021 |DATE_PERIOD |
|scope 3 emissions |ENVIRONMENTAL_KPI |
|sold |ECONOMIC_ACTION |
|products |SOCIAL_KPI |
|emissions |ENVIRONMENTAL_KPI |
|377 million |AMOUNT |
|tonnes of CO2e |ENVIRONMENTAL_UNIT|
|2017 |DATE_PERIOD |
|408 million |AMOUNT |
|tonnes of CO2e |ENVIRONMENTAL_UNIT|
|2021 |DATE_PERIOD |
+----------------------+------------------+
Model Information
| Model Name: | finner_responsibility_reports |
| Compatibility: | Finance NLP 1.0.0+ |
| License: | Licensed |
| Edition: | Official |
| Input Labels: | [document, token] |
| Output Labels: | [ner] |
| Language: | en |
| Size: | 406.6 MB |
| Case sensitive: | false |
| Max sentence length: | 512 |
References
In-house annotations on Responsibility and ESG Reports
Benchmarking
label precision recall f1-score support
AGE 0.86 0.84 0.85 37
AMOUNT 0.93 0.96 0.95 1254
COUNTABLE_ITEM 0.87 0.86 0.87 212
DATE_PERIOD 0.90 0.93 0.92 925
ECONOMIC_ACTION 0.83 0.85 0.84 61
ECONOMIC_KPI 0.78 0.83 0.80 223
ENVIRONMENTAL_ACTION 0.84 0.84 0.84 332
ENVIRONMENTAL_KPI 0.79 0.86 0.82 948
ENVIRONMENTAL_UNIT 0.91 0.90 0.91 484
ESG_ROLE 0.76 0.81 0.79 139
FACILITY_PLACE 0.70 0.88 0.78 154
ISO 0.68 0.81 0.74 32
PERCENTAGE 0.98 1.00 0.99 706
PROFESSIONAL_GROUP 0.88 0.95 0.91 419
RELATIVE_METRIC 0.92 0.94 0.93 141
SOCIAL_ACTION 0.83 0.81 0.82 262
SOCIAL_KPI 0.82 0.84 0.83 480
TARGET_GROUP 0.84 0.92 0.88 257
TARGET_GROUP_BUSINESS 0.93 0.98 0.96 44
WASTE 0.80 0.77 0.79 106
micro-avg 0.87 0.91 0.89 7216
macro-avg 0.84 0.88 0.86 7216
weighted-avg 0.87 0.91 0.89 7216