Weighted Euclidean Distancetype
Post
status
Published
date
Jul 2, 2021
slug
summary
tags
Mathematics
category
Mathematics
icon
password
Calculating Product Similarity Using Weighted Euclidean Distance
Suppose we have two products, Product A and Product B, and each has the following features:
Feature Type | Feature Name | Product A | Product B | Weight (wᵢ) |
Image Feature | Image Vector | 0.5 | 0.7 | 0.4 |
Text Feature | Title Vector | 0.8 | 0.6 | 0.3 |
Categorical | Brand (One-hot) | 1.0 | 0.0 | 0.2 |
Categorical | Color (One-hot) | 1.0 | 1.0 | 0.1 |
✅ Step-by-Step Calculation
We'll use the formula:
1. Image Feature:
2. Title/Text Feature:
3. Brand (One-hot):
4. Color (One-hot):
Final Distance:
Interpretation:
- The smaller the distance, the more similar the products are.
- In this example, the brand difference contributed the most to the distance due to its high weight and binary difference.
- The image and text also influenced similarity moderately.
- Color was identical, so it had no effect on the distance.
- Author:NotionNext
- URL:https://tangly1024.com/article/1c6d698f-3512-810f-84be-f889d024dd36
- Copyright:All articles in this blog, except for special statements, adopt BY-NC-SA agreement. Please indicate the source!
