This repository contains utils to construct vector models from navigational paths, e.g.
from the WikiGame, Wiki Clickstream or WikiSpeedia.
The code in this repository is mainly based on Apache Spark and Python.
For methodological details, kindly refer to the following publications:
- West et al: “Wikispeedia: An Online Game for Inferring Semantic Distances between Concepts”
- Singer et al: “Computing Semantic Relatedness from Human Navigation: A Case Study on Wikipedia”
- Niebler et al: “Extracting Semantic Relatedness from Unconstrained Navigation on Wikipedia”
- Dallmann et al: “Extracting Semantic Relatedness from Random Walks on Wikipedia: Copmaring Learning and Counting Methods”
I have created a short notebook to demonstrate how to use the presented utils. It can be found either in the
repository or beautifully
rendered in Markdown on this page.