Description
This is a sm
(small) version of a financial model trained on Broker Reports to detect financial entities (NER model).
Predicted Entities
LIABILITY_INCREASE
, REVENUE_INCREASE
, ASSET_DECREASE
, AMOUNT
, TICKER
, TARGET_PRICE
, ORG
, DATE
, LIABILITY_DECREASE
, LIABILITY
, CFO_INCREASE
, ASSET_INCREASE
, LOSS
, CMP
, ASSET
, CF_DECREASE
, EXPENSE
, CF
, PAD
, CFO
, FCF
, PROFIT_INCREASE
, REVENUE_DECLINE
, CF_INCREASE
, PERCENTAGE
, RATING
, STOCKHOLDERS_EQUITY
, PROFIT_DECLINE
, PROFIT
, CURRENCY
, FISCAL_YEAR
, EXPENSE_INCREASE
, EXPENSE_DECREASE
, REVENUE
How to use
documentAssembler = nlp.DocumentAssembler()\
.setInputCol("text")\
.setOutputCol("document")
tokenizer = nlp.Tokenizer()\
.setInputCols("document")\
.setOutputCol("token")
tokenClassifier = finance.BertForTokenClassification.pretrained("finner_broker_reports_specific_sm","en","finance/models")\
.setInputCols("token", "document")\
.setOutputCol("label")\
.setCaseSensitive(True)
converter = finance.NerConverterInternal()\
.setInputCols(["document", "token", "label"])\
.setOutputCol("ner_span")
pipeline = nlp.Pipeline(
stages=[
documentAssembler,
tokenizer,
tokenClassifier,
converter
]
)
Results
+-----------------+----------+
|chunk |entity |
+-----------------+----------+
|revenue |REVENUE |
|$ |CURRENCY |
|1.7 billion |AMOUNT |
|net profit margin|PROFIT |
|13 |PERCENTAGE|
|net debt |LIABILITY |
|Rs |CURRENCY |
|6.62bn |AMOUNT |
+-----------------+----------+
Model Information
Model Name: | finner_broker_reports_specific_sm |
Compatibility: | Finance NLP 1.0.0+ |
License: | Licensed |
Edition: | Official |
Input Labels: | [sentence, token] |
Output Labels: | [ner] |
Language: | en |
Size: | 400.9 MB |
Case sensitive: | true |
Max sentence length: | 128 |
References
In-house annotated dataset
Benchmarking
labels precision recall f1-score support
B-REVENUE_INCREASE 0.69 0.78 0.73 126
I-ASSET_DECREASE 0.80 0.87 0.83 23
B-ASSET 0.84 0.82 0.83 50
B-REVENUE_DECLINE 0.81 0.79 0.80 28
I-STOCKHOLDERS_EQUITY 1.00 0.96 0.98 56
B-CF 0.77 0.94 0.85 18
I-LOSS 0.92 0.92 0.92 25
I-REVENUE 0.28 0.26 0.27 19
I-PROFIT_DECLINE 0.80 0.94 0.86 17
I-PROFIT 0.91 0.94 0.93 249
I-EXPENSE_DECREASE 0.77 1.00 0.87 10
I-REVENUE_INCREASE 0.58 0.68 0.63 56
B-TICKER 0.65 0.81 0.72 73
B-EXPENSE 0.76 0.90 0.83 63
I-CF_DECREASE 0.90 1.00 0.95 38
B-RATING 0.93 0.99 0.96 536
B-FCF 1.00 1.00 1.00 18
B-CF_INCREASE 1.00 1.00 1.00 20
B-CURRENCY 0.98 1.00 0.99 936
I-ASSET_INCREASE 0.76 1.00 0.86 16
B-CFO 0.96 1.00 0.98 22
I-LIABILITY 0.97 0.82 0.89 38
B-LOSS 0.85 0.94 0.89 31
I-LIABILITY_DECREASE 0.44 0.67 0.53 12
B-PROFIT_INCREASE 0.81 0.77 0.79 173
B-EXPENSE_DECREASE 0.90 1.00 0.95 26
B-CFO_INCREASE 0.92 0.96 0.94 25
B-AMOUNT 0.97 1.00 0.98 1999
B-TARGET_PRICE 0.97 0.98 0.97 171
B-PROFIT 0.85 0.92 0.88 437
I-AMOUNT 0.97 0.97 0.97 464
B-ASSET_INCREASE 0.69 0.78 0.73 23
I-EXPENSE_INCREASE 0.88 0.79 0.84 29
B-LIABILITY_DECREASE 0.85 0.88 0.86 40
B-PROFIT_DECLINE 0.90 0.85 0.87 53
B-STOCKHOLDERS_EQUITY 1.00 0.93 0.96 27
I-EXPENSE 0.91 0.87 0.89 61
B-PERCENTAGE 0.95 0.99 0.97 1133
I-ASSET 0.97 0.73 0.83 44
I-PERCENTAGE 0.33 1.00 0.50 2
I-TARGET_PRICE 0.94 1.00 0.97 119
B-CMP 0.89 1.00 0.94 33
I-PROFIT_INCREASE 0.75 0.75 0.75 48
I-DATE 0.40 0.57 0.47 14
B-REVENUE 0.78 0.83 0.80 128
B-ASSET_DECREASE 0.87 0.91 0.89 22
I-CF_INCREASE 1.00 1.00 1.00 42
I-FISCAL_YEAR 0.98 0.89 0.93 110
I-CFO 0.97 1.00 0.99 75
I-FCF 1.00 1.00 1.00 32
B-ORG 0.95 0.98 0.96 1310
I-ORG 0.95 0.98 0.97 1005
B-EXPENSE_INCREASE 1.00 0.68 0.81 34
I-REVENUE_DECLINE 0.67 0.38 0.48 21
I-RATING 0.00 0.00 0.00 1
B-DATE 0.98 0.99 0.98 358
I-CMP 0.97 1.00 0.98 29
B-LIABILITY_INCREASE 1.00 0.93 0.96 41
B-FISCAL_YEAR 0.96 0.89 0.93 28
I-CURRENCY 0.96 1.00 0.98 72
I-CFO_INCREASE 0.89 0.97 0.93 72
B-CF_DECREASE 0.96 1.00 0.98 23
I-TICKER 1.00 0.40 0.57 5
I-LIABILITY_INCREASE 1.00 0.94 0.97 31
B-LIABILITY 0.96 0.92 0.94 26
I-CF 0.80 0.85 0.83 39
micro-avg 0.93 0.96 0.95 10905
macro-avg 0.85 0.87 0.85 10905
weighted-avg 0.93 0.96 0.95 10905