Nosto Search' algorithms provide mechanisms to detect different queries that are based on the same root word (e.g. pluralised forms like "dress" -> "dresses"). This allows the algorithm to return matching results even if the query is not exactly the product data. This is called stemmed matching.
If the algorithm detects queries that are similar to words in the product data but are not sharing the same root word, this is called "Fuzzy Search" (e.g. query is "dross", whereas product data contains the word "dress"). Fuzzy Search is only triggered when the regular search (including exact, stemmed and prefix matches) is not returning any results to provide the user with similar results (that might derive for example from typos).
Fuzzy Search might get applied in different ways (e.g. "dross" -> "dress" (fuzzy exact match) or "dross -> dresses" (fuzzy stemmed match).
In Score Insights you can learn from the Fuzzy Search Indicator if a Fuzzy Search got applied. This way you know that the regular search for this particular query didn't return any results and the algorithm looked for some syntactical similar words in your product data.