Semi-Metric Behavior in Document Networks and Its Application to Recommendation Systems

Title: Semi-Metric Behavior in Document Networks and Its Application to Recommendation Systems

Abstract: This research explores the concept of semi-metric distance graphs and their application to recommendation systems. It aims to understand the behavior of these graphs and develop a model to quantify the importance of latent associations in a Document Network (DN). The study also focuses on combining evidence from different distance graphs to improve recommendation systems.

Research Question: How can semi-metric distance graphs be used to improve recommendation systems?

Methodology: The study uses a multi-agent algorithm called TalkMine, which integrates evidence from different distance graphs. It introduces ratios to measure semi-metric behavior and computes these ratios for various DNs like digital libraries and web sites. The study also proposes a model based on semi-metric behavior to quantify the amount of important latent associations in a DN.

Results: The study finds that the ratios measured from semi-metric distance graphs are useful in identifying implicit associations. The proposed model successfully quantifies the importance of latent associations in a DN. The algorithm developed for TalkMine effectively combines evidence from different distance graphs, improving the recommendation system's performance.

Implications: The research has significant implications for the field of recommendation systems. It provides a new approach to understanding and improving recommendation systems by using semi-metric distance graphs. The study also offers a practical solution for combining evidence from different distance graphs, which can lead to more accurate and personalized recommendations.

Link to Article: https://arxiv.org/abs/0309013v1 Authors: arXiv ID: 0309013v1