Attribute filtering / Selection
Aino AI allows you to filter your GIS datasets using natural-language queries. The AI automatically identifies the most relevant attribute columns and applies filters based on your request.
1. Filter by Existence / Non-Existence of an Attribute Value
Use natural language to find features where a specific attribute exists or is missing.
Example prompt:
Show schools that have a website in "Schools"
Result:
Returns all records in the Schools layer where the website field exists and is not empty.
2. Filter by Attribute Value (Partial Match)
Aino AI can interpret and match partial strings or perform fuzzy searches across relevant fields. It predicts which attributes are most likely to contain the requested value.
Example prompt:
Result:
Finds all records where attributes such as category, type, or description contain the word “cosmetology” or similar.
How it works: When deciding to filter by a string, Aino does not use strict equality. Instead, it calculates the Levenshtein distance between the input value and each column. A dynamic similarity threshold is applied based on the length of the input string: longer strings have a stricter threshold.
The threshold calculation works as follows:
Clean the string, keeping only English letters and digits.
Measure the length of the cleaned string.
Apply a monotonically decreasing function (inverse proportionality with an asymptote at 1/3) to determine the similarity threshold.
Longer strings → lower threshold → more precise matching.
Shorter strings → higher threshold → more flexible matching.
3. Filter by condition (numeric / range / comparison)
Aino AI supports numeric filtering using comparisons, ranges, and inequalities (>, <, >=, <=, between, etc.).
Example prompt:
Result:
Returns all records in the Buildings_of_Monaco layer where the levels attribute is greater than 3.
We recommend using a forward slash (/) before each dataset name you want to filter.
For example:
Show schools that have a website in /Schools
Last updated