Measuring Research Interest Similarity with Transition Probabilities

2024-09-26

arXiv.org

AI Summary

This paper addresses the challenge of accurately quantifying similarity between scholarly papers and authors’ research interests in literature retrieval. We propose a Transfer Probability (TP) metric based on random walks over citation networks, formalizing information retrieval as probabilistic transitions on such networks. TP requires no predefined disciplinary categories, manual annotations, or hard clustering, and directly yields continuous, interpretable, fine-grained similarity scores. Compared to embedding methods like Node2Vec, TP achieves superior performance in capturing macro-level disciplinary structure and modeling author-level research interests. An open-source Python toolkit implements multiple TP variants, demonstrating its effectiveness and robustness across diverse scenarios. To our knowledge, this is the first work to systematically integrate retrieval principles into citation-network-based similarity modeling, establishing a novel paradigm for cross-domain research interest analysis.

Technology Category

• Reasoning under Uncertainty: Relational Probabilistic Models
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Knowledge Representation and Reasoning: Preferences

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Querying, indexing, and retrieval in Web-related graphs
• Web Mining and Content Analysis: Content-based information diffusion

Abstract

We propose a method to measure the similarity of papers and authors by simulating a literature search procedure on citation networks, which is an information retrieval inspired conceptualization of similarity. This transition probability (TP) based approach does not require a curated classification system, avoids clustering complications, and provides a continuous measure of similarity. We perform testing scenarios to explore several versions of the general TP concept and the Node2vec machine-learning technique. We found that TP measures outperform Node2vec in mapping the macroscopic structure of fields. The paper provides a general discussion of how to implement TP similarity measurement, with a particular focus on how to utilize publication-level information to approximate the research interest similarity of individual scientists. This paper is accompanied by a Python package capable of calculating all the tested metrics.

Problem

• Measuring research similarity through citation network transition probabilities
• Comparing similarity metrics for paper classification and co-author prediction
• Developing continuous symmetric measures for any citation network structure

Innovation

• Transition probability models citation network browsing
• Continuous symmetric measures work on any citation network
• Outperforms PageRank and Node2vec in classification tasks