Search code examples
pythoneuclidean-distanceorange

Query Relaxation - Orange

The objective here is to relax a failing query by substituting it with the closest association rule from the dataset in picture.

My idea is to find the euclidean distances between each association rule and the query, sort the distances and select the rule with the least distance.

I am facing a problem in finding distances between the rules and the query. I'm not sure if weights have to be assigned to the items in each rule. If so, how do I assign these weights?

I'm working in Orange and I don't have much Python (or any programming) experience. I need all the help I can get.

Here's the Orange Dashboard of what I've done until now.

Orange Dashboard

When I try to connect the example distance node with the assoc rules explorer node, I get a window given below, which doesn't allow me to connect association rules to the data

enter image description here

Solution

  • Orange can only compute distances between examples, not between rules. You can't do that in canvas, you'll need some Python programming for this.