The paper presents an interactive web platform that allows users to compute personalized relevance scores on a set of example graphs, including Wikilink networks from different language editions, networks of interaction on Twitter about specific topics, and networks of co-purchased products from Amazon. The platform showcases several algorithms for computing personalized relevance, including:
The platform enables two main use cases: (1) algorithm comparison, where users can compare the results obtained with different algorithms, and (2) dataset comparison, where users can explore and gain insights into a dataset and compare it with others.
The system architecture consists of four main components: a datastore, an API gateway, computational nodes, and a web user interface. The web interface allows users to select the dataset, algorithm, and parameters to execute, and visualizes the results.
The paper showcases the use of the platform through concrete examples, highlighting how Cyclerank can produce more relevant results compared to the other algorithms, particularly in addressing the issue of over-representation of popular nodes.
To Another Language
from source content
arxiv.org
Önemli Bilgiler Şuradan Elde Edildi
by Luca Cavalca... : arxiv.org 05-06-2024
https://arxiv.org/pdf/2405.02261.pdfDaha Derin Sorular