# Lookup table is supposed to classify entities, but does it influence intent prediction? **URL:** \ **Category:** Rasa Open Source\ **Created:** [March 7, 2020, 10:59am UTC](https://forum.rasa.com/t/lookup-table-is-supposed-to-classify-entities-but-does-it-influence-intent-prediction/25776 "2020-03-07T10:59:47Z")\ **Posts on this page:** 1\ **Showing post:** 3
**Author:** ![slcheungcasado](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/slcheungcasado/32/7865_2.png) [@slcheungcasado](https://forum.rasa.com/u/slcheungcasado)\ **Post date:** [March 16, 2020, 2:24pm UTC](https://forum.rasa.com/t/lookup-table-is-supposed-to-classify-entities-but-does-it-influence-intent-prediction/25776/3 "2020-03-16T14:24:40Z")
I’ll leave [this](https://blog.rasa.com/rasa-nlu-in-depth-part-2-entity-recognition/?_ga=2.126697551.358823277.1584348790-1684436698.1582097539) link here as additional reference. This passage mentions something important about the entity extraction aspect of lookup tables. > Regular expressions and lookup tables are adding additional features to _ner\_crf_ which mark whether a word was matched by a regular expression or lookup table entry. As it is one feature of many, the component _ner\_crf_ can still ignore an entity although it was matched, however in general _ner\_crf_ develops a bias for these features. **Note that this can also stop the conditional random field from generalizing: if all entity examples in your training data are matched by a regular expression, the conditional random field will learn to focus on the regular expression feature and ignore the other features.** If you then have a message with a certain entity which is not matched by the regular expression, _ner\_crf_ will probably not be able to detect it. **Especially the use of lookup tables makes _ner\_crf_ prone for overfitting.** --- _[View the full topic](https://forum.rasa.com/t/lookup-table-is-supposed-to-classify-entities-but-does-it-influence-intent-prediction/25776)._