# Intentions and confidence

**URL:** <https://forum.rasa.com/t/intentions-and-confidence/43026>\
**Category:** Rasa Open Source\
**Created:** [April 29, 2021, 1:58pm UTC](https://forum.rasa.com/t/intentions-and-confidence/43026 "2021-04-29T13:58:17Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![joggerjoel](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/joggerjoel/32/14583_2.png) [@joggerjoel](https://forum.rasa.com/u/joggerjoel)\
**Post date:** [April 29, 2021, 1:58pm UTC](https://forum.rasa.com/t/intentions-and-confidence/43026/1 "2021-04-29T13:58:17Z")

</div>

Is there any way to reduce the confidence on incoming words if I know that the ML is super inaccurate?

input\> can’t right now

message\_intent: c\_features message\_confidence: .6839

```auto
 intent: c_features
  examples: |
    - Features
    - C features
    - What are the features of C?
    - Advantages of C
    - C Programming Features

```

I don’t actually understand how the confidence is so high when in fact no words even closely match… but, nevertheless; how can I for manually edit this confidence in python?

---

<div class="post-metadata">

**Author:** ![joggerjoel](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/joggerjoel/32/14583_2.png) [@joggerjoel](https://forum.rasa.com/u/joggerjoel)\
**Post date:** [April 29, 2021, 2:19pm UTC](https://forum.rasa.com/t/intentions-and-confidence/43026/2 "2021-04-29T14:19:43Z")

</div>

I don’t think there’s anything out-of-box, so I’ll handle it with some custom Python code… But I am curious why the confidence is so high for this specific scenario.

---

<div class="post-metadata">

**Author:** ![tiny\_imod](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/tiny_imod/32/10621_2.png) [@tiny\_imod](https://forum.rasa.com/u/tiny_imod)\
**Post date:** [April 30, 2021, 8:52am UTC](https://forum.rasa.com/t/intentions-and-confidence/43026/3 "2021-04-30T08:52:51Z")

</div>

Hello Joel,

Thank you for your question! I would recommend to try to train your bot so it does give you the correct intent prediction instead of changing the confidence by brute force (with custom Python code).

Some tips on debugging :

- try to run with `--debug` flag to see what NLU is doing (as already mentioned [here](http://forum.rasa.com/t/adding-new-intent-and-action-to-starter-pack-is-not-working/6030/2) also)
- try to add some more examples under the `c_features` intent (10-20)
- align your intent classifier with the rest of your pipeline (`config.yml`) or share it and I can also have a look
- make sure that your intents are not too specific and each one represents a broad intent that the user is trying to do

The last two are also mentioned in this [video](https://www.youtube.com/watch?v=JOf6CNJUzEo) from Rachael in section `common errors` (min 4:29).

Also, some resources that might be useful for you are the following :

- [Here](https://blog.rasa.com/rasa-nlu-in-depth-part-1-intent-classification/) you can find a very insightful Rasa blog post about intent classification
- [Here](https://blog.rasa.com/intents-entities-understanding-the-rasa-nlu-pipeline/) is a very good Rasa blog post that can give you more insights on your NLU pipeline
- Confidence explained by Rachael [here](https://www.youtube.com/watch?v=7u8hrogfPZI)
- This [post](http://forum.rasa.com/t/rasa-train-taking-lot-of-time/31367/11) gives an insightful explanation on accuracy