Sahar Millis
Reputation: 907
Using BERT for extracting Product Features
How can I use BERT to extract product features from text?
For example, how to extract the ids from the next emails - WSK30015675, KTXFFC156, 4569TT11DRE
Hello John, Please order 15 dresses ID WSK30015675 from our storage. Thanks.
Hello Dan, this week we need to make sure we have KTXFFC156. And please make sure to order more of 4569TT11DRE. tnx :)
Notes:
- No regex
- The ids are unknown
- There is enough tagged data (emails >> ids)
- The ids are combinations of numbers and letter in different length
Upvotes: 1
Views: 1019
Answers (1)
Omk
Reputation: 330
you need to detect which words are product features so you should use this archirecture
this architecture implemented in similar tasks like these:
https://gab41.lab41.org/how-to-fine-tune-bert-for-named-entity-recognition-2257b5e5ce7e
https://towardsdatascience.com/named-entity-recognition-ner-with-bert-in-spark-nlp-874df20d1d77
Upvotes: 1
Related Questions
- Using trained BERT Model and Data Preprocessing
- Extracting Features from BertForSequenceClassification
- Trying to detect products from text while using a dictionary
- Extracting word features from BERT model
- BERT NER Python
- Extracting Product Attribute/Features from text
- How to Extract Features from Text based on Fine-Tuned BERT Model
- Use BERT for feature extraction of a unique word
- Extract brand and product category from consumer product manuals
- Extract product name from english text