====AI vs Fuzzy Matching====

If you're trying to use your captured data to suggest setpoints, system setups, recipes etc, there are two main options for how you can search for the values you need.

**Fuzzy Matching** uses smart search to find a time where you've made a similar product in the past.

**AI Modelling** uses a learning artificial intelligence to suggest what your settings should be.

===Advantages of Fuzzy Matching===

Compared to AI results, Fuzzy Matching results...

  * Are guaranteed to be reasonable and to have worked before,
  * Can indicate if matches are exact or if there are differences (highlighting that the match is poor or that your situation is new),
  * Are traceable - it tells you exactly **which** capture it matched with, and when it happened.

===Advantages of AI===

However, compared to Fuzzy Matching, AI...

  * Is much better at predicting values in new configurations (within limits),
  * Is able to compensate for a wide range of factors with much less information,
  * Is able to run quickly, even with very large amounts of data to search
  * Is more suitable for large 'problem spaces' (ie. situations with many analogue inputs).