RASA Single Character Entity Recognition Problem

I am working on a bot where some of my entity can be single character. For example under the model.txt lookup table i mentioned all the related model names like R1s , F11s, A34Mp.

In nlu.md file i mentioned quries like

  • from where i can buy R1s
  • i need to buy A34Mp

But some times users ask queries like “i need to buy A 34 MP” or “i want to buy R 1 s”.

In this case when user type the entity name in single character form , bot does not extract the entity name/value.

I also mention the name in synonyms in nlu.md file like

synonym:R1S

  • R 1 S
  • R1 S
  • R 1S

But still rasa fails to extract the entity. What do i need to in nlu.md or config.yml or any thing else. Please help if there is any way to solve this problem.

Thanks

1 Like

Use this pipeline: language: en

pipeline:

  • name: “SpacyNLP” model: “en_core_web_md”
  • name: “SpacyTokenizer”
  • name: “RegexFeaturizer”
  • name: “LexicalSyntacticFeaturizer”
  • name: “CountVectorsFeaturizer”
  • name: “CountVectorsFeaturizer” analyzer: “char_wb” min_ngram: 1 max_ngram: 4
  • name: “DIETClassifier” epochs: 100
  • name: “EntitySynonymMapper”