Description
This model is a Named Entity Recognition (NER) model focused on financial numeric items. It identifies 13 numeric financial entities from diverse 10-Q and 10-K reports. These entities are annotated using eXtensible Business Reporting Language (XBRL) tags. The annotation process primarily targets numerical tokens, and the context plays a crucial role in accurately assigning the appropriate entity type from the 139 most common financial entities available in the dataset.
Predicted Entities
Goodwill, AcquiredFiniteLivedIntangibleAssetsWeightedAverageUsefulLife, OperatingLeaseRightOfUseAsset, PropertyPlantAndEquipmentUsefulLife, AreaOfRealEstateProperty, NumberOfRealEstateProperties, CashAndCashEquivalentsFairValueDisclosure, FiniteLivedIntangibleAssetUsefulLife, EquityMethodInvestments, BusinessAcquisitionPercentageOfVotingInterestsAcquired, OperatingLossCarryforwards, DerivativeNotionalAmount, EquityMethodInvestmentOwnershipPercentage
How to use
documentAssembler = nlp.DocumentAssembler() \
.setInputCol("text") \
.setOutputCol("document")
sentence = nlp.SentenceDetector() \
.setInputCols(["document"]) \
.setOutputCol("sentence")
tokenizer = nlp.Tokenizer()\
.setInputCols(["sentence"])\
.setOutputCol("token")\
.setContextChars(['.', ',', ';', ':', '!', '?', '*', '-', '(', ')', '”', '’', '$','€'])
embeddings = nlp.BertEmbeddings.pretrained("bert_embeddings_sec_bert_base","en") \
.setInputCols(["document", "token"]) \
.setOutputCol("embeddings")\
.setMaxSentenceLength(512)
nerTagger = finance.NerModel.pretrained('finner_10q_xbrl_lg_asset', 'en', 'finance/models')\
.setInputCols(["sentence", "token", "embeddings"])\
.setOutputCol("ner")
pipeline = nlp.Pipeline(stages=[documentAssembler,
sentence,
tokenizer,
embeddings,
nerTagger
])
text = "As of May 31 , 2016 , total assets include $ 645.1 million related to consolidated VIEs of which $ 8.2 million is included in Lennar Homebuilding cash and cash equivalents , $ 0.1 million in Lennar Homebuilding receivables , net , $ 6.2 million in Lennar Homebuilding finished homes and construction in progress , $ 158.8 million in Lennar Homebuilding land and land under development , $ 134.5 million in Lennar Homebuilding consolidated inventory not owned , $ 4.5 million in Lennar Homebuilding investments in unconsolidated entities , $ 21.4 million in Lennar Homebuilding other assets , $ 280.0 million in Rialto assets and $ 31.4 million in Lennar Multifamily assets ."
df = spark.createDataFrame([[text]]).toDF("text")
fit = pipeline.fit(df)
result = fit.transform(df)
result_df = result.select(F.explode(F.arrays_zip(result.token.result,result.ner.result, result.ner.metadata)).alias("cols"))\
.select(F.expr("cols['0']").alias("token"),\
F.expr("cols['1']").alias("ner_label"),\
F.expr("cols['2']['confidence']").alias("confidence"))
result_df.show(50, truncate=100)
Results
+--------------+-------------------------+----------+
|token |ner_label |confidence|
+--------------+-------------------------+----------+
|As |O |1.0 |
|of |O |1.0 |
|May |O |1.0 |
|31 |O |1.0 |
|, |O |1.0 |
|2016 |O |1.0 |
|, |O |1.0 |
|total |O |1.0 |
|assets |O |1.0 |
|include |O |1.0 |
|$ |O |1.0 |
|645.1 |O |0.9998 |
|million |O |1.0 |
|related |O |1.0 |
|to |O |1.0 |
|consolidated |O |1.0 |
|VIEs |O |1.0 |
|of |O |1.0 |
|which |O |1.0 |
|$ |O |1.0 |
|8.2 |O |1.0 |
|million |O |1.0 |
|is |O |1.0 |
|included |O |1.0 |
|in |O |1.0 |
|Lennar |O |1.0 |
|Homebuilding |O |1.0 |
|cash |O |1.0 |
|and |O |1.0 |
|cash |O |1.0 |
|equivalents |O |1.0 |
|, |O |1.0 |
|$ |O |1.0 |
|0.1 |O |0.9995 |
|million |O |1.0 |
|in |O |1.0 |
|Lennar |O |1.0 |
|Homebuilding |O |1.0 |
|receivables |O |1.0 |
|, |O |1.0 |
|net |O |1.0 |
|, |O |1.0 |
|$ |O |1.0 |
|6.2 |O |0.9991 |
|million |O |1.0 |
|in |O |1.0 |
|Lennar |O |1.0 |
|Homebuilding |O |1.0 |
|finished |O |1.0 |
|homes |O |1.0 |
|and |O |1.0 |
|construction |O |1.0 |
|in |O |1.0 |
|progress |O |1.0 |
|, |O |1.0 |
|$ |O |1.0 |
|158.8 |O |0.9935 |
|million |O |1.0 |
|in |O |1.0 |
|Lennar |O |1.0 |
|Homebuilding |O |1.0 |
|land |O |1.0 |
|and |O |1.0 |
|land |O |1.0 |
|under |O |1.0 |
|development |O |1.0 |
|, |O |1.0 |
|$ |O |1.0 |
|134.5 |O |0.9959 |
|million |O |1.0 |
|in |O |1.0 |
|Lennar |O |1.0 |
|Homebuilding |O |1.0 |
|consolidated |O |1.0 |
|inventory |O |0.9997 |
|not |O |0.9999 |
|owned |O |1.0 |
|, |O |1.0 |
|$ |O |1.0 |
|4.5 |B-EquityMethodInvestments|0.9934 |
|million |O |1.0 |
|in |O |1.0 |
|Lennar |O |1.0 |
|Homebuilding |O |1.0 |
|investments |O |1.0 |
|in |O |1.0 |
|unconsolidated|O |1.0 |
|entities |O |1.0 |
|, |O |1.0 |
|$ |O |1.0 |
|21.4 |O |0.9917 |
|million |O |1.0 |
|in |O |1.0 |
|Lennar |O |1.0 |
|Homebuilding |O |1.0 |
|other |O |1.0 |
|assets |O |1.0 |
|, |O |1.0 |
|$ |O |1.0 |
|280.0 |O |0.9985 |
+--------------+-------------------------+----------+
Model Information
| Model Name: | finner_10q_xbrl_lg_asset |
| Compatibility: | Finance NLP 1.0.0+ |
| License: | Licensed |
| Edition: | Official |
| Input Labels: | [sentence, token, embeddings] |
| Output Labels: | [ner] |
| Language: | en |
| Size: | 16.4 MB |
References
An in-house modified version of https://huggingface.co/datasets/nlpaueb/finer-139, re-splited and filtered to focus on sentences with bigger density of tags.
Benchmarking
label precision recall f1-score support
B-AcquiredFiniteLivedIntangibleAssetsWeightedAverageUsefulLife 0.6955 0.8158 0.7509 266
B-AreaOfRealEstateProperty 0.9558 0.9251 0.9402 187
B-BusinessAcquisitionPercentageOfVotingInterestsAcquired 0.9490 0.5820 0.7215 256
B-CashAndCashEquivalentsFairValueDisclosure 1.0000 0.9827 0.9913 231
B-DerivativeNotionalAmount 0.9829 0.9844 0.9836 641
B-EquityMethodInvestmentOwnershipPercentage 0.8688 0.9851 0.9233 874
B-EquityMethodInvestments 0.8880 0.9715 0.9279 351
B-FiniteLivedIntangibleAssetUsefulLife 0.8199 0.7629 0.7904 388
B-Goodwill 0.9910 0.9379 0.9637 467
B-NumberOfRealEstateProperties 0.8981 0.9831 0.9387 296
B-OperatingLeaseRightOfUseAsset 0.9861 0.8903 0.9357 319
B-OperatingLossCarryforwards 0.9819 0.9056 0.9422 180
B-PropertyPlantAndEquipmentUsefulLife 0.9641 0.9335 0.9486 316
I-AcquiredFiniteLivedIntangibleAssetsWeightedAverageUsefulLife 0.8140 0.5833 0.6796 60
I-AreaOfRealEstateProperty 0.0000 0.0000 0.0000 1
I-FiniteLivedIntangibleAssetUsefulLife 0.6869 0.7556 0.7196 90
I-NumberOfRealEstateProperties 1.0000 0.5000 0.6667 2
I-PropertyPlantAndEquipmentUsefulLife 0.9524 0.7595 0.8451 79
O 0.9987 0.9989 0.9988 123678
accuracy - - 0.9952 128682
macro-avg 0.8649 0.8030 0.8246 128682
weighted-avg 0.9954 0.9952 0.9951 128682