Estimating The MBTI Of Twitter Accounts With Graph Neural Networks Over Neo4j

Sep 17, 2026

Uploaded by Skyler

Intelligent agents serve as flexible tools for efficiently mining large graphs to extract heterogeneous knowledge. Using Twitter as a central case, account behavior can be inferred from structural and functional signals such as original multimedia posting and retweeting, supported by hashtag emotional polarity and reactions to events. The work extends from emotional reactions to integrated personality estimation via the MBTI taxonomy, using a graph neural network and an intelligent account visiting policy with preferential attachment and optional forgetting.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages10

Reading Time25 min

AccessFree to download

Abstract

1. Introduction

Estimating The MBTI Of Twitter Accounts With Graph Neural Networks Over Neo4j Georgios Drakopoulos1,∗ , Eleanna Kafeza2 1 Department of Informatics, Ionian University, Tsirigoti Sq. 7, Kerkyra 49100, Hellas 2 College of Interdisciplinary Studies, Zayed University, UAE Abstract Intelligent agents are indispensable and flexible autonomous tools for efficiently mining large graphs for heterogeneous knowledge. Twitter is a prime case in point with structural and functional attributes such as original multimedia content posting and retweeting revealing important affective information about accounts. Additionally, this can be facilitated by including hashtag emotional polarity and reactions to political, social, or even historical events. Further insight can be gained by moving one step forward from individual emotional reactions to integrated personality estimations such as the MBTI taxonomy. An intelligent agent has been developed with a stochastic account visiting policy based on preferential attachment, an optional evolving forgetting factor for penalizing vertices appearing too frequently, and the capability to yield an MBTI estimate based on a graph neural network. The results indicate the superior performance of the proposed heuristic based on evaluation criteria including community size distribution and hashtag coherency. Keywords Intelligent agents, random walks, graph neural networks, affective communities, emotional polarity, personality taxonomies, MBTI, preferential attachment, Twitter, Neo4j, py2neo, PyTorch 1. Introduction Intelligent agents (IAs) are autonomous digital entities with extensive capabilities for maintaining and ensuring the smooth operation of massive infrastructure, mostly networks of various types. Depending mainly on their technology and operating principles, an IA may have extended command and control capabilities while requiring only the initial programming in order to properly function, excluding of course any sort of necessary local input. Recently IA has advanced beyond a fixed set of rules to integrating various level of machine learning (ML) capabilities, provided sufficient computing power is available. The inclusion of neural networks architectures such as graph neural networks (GNNs) which take full advantage of the local network topology constitutes a major addition. Perhaps the most well-known representation of such an agent in pop culture, albeit possessing far more capabilities than those of its contemporary counterparts, is that of the iconic agent Smith from The Matrix 1. Twitter mining analytics provide a significant opportunity to various organizations to improve both deci- CIKM’22: 31st ACM International Conference on Information and Knowledge Management (companion volume), October 17 –21, 2022, Atlanta, GA ∗Corresponding author. [Envelope-Open](Envelope-Open c16drak@ionio.gr)[ c16drak@ionio.gr](Envelope-Open c16drak@ionio.gr) (G. Drakopoulos); [eleana.kafeza@zu.ac.ae](eleana.kafeza@zu.ac.ae)[ ](eleana.kafeza@zu.ac.ae)(E. Kafeza) Orcid 0000-0002-0975-1877 (G. Drakopoulos); 0000-0001-9565-2375 (E. Kafeza) © 2022 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4 .0 International (CC BY 4 .0) . CEURWorkshopProceedings CEUR Workshop Proceedings ([CEUR-WS.org](CEUR-WS.org)) 1[https://www.imdb.com/title/tt0133093](https://www.imdb.com/title/tt0133093) [http://ceur-ws.org](http://ceur-ws.org) [ISSN 1613-0073](ISSN 1613-0073) sion making processes and marketing performance. This is achieved as businesses are allowed through Twitter mining to gain invaluable insights on the dynamics and collective behavior of their customer base or any other online target group for that matter. In turn this yields more accurate predictions of key factors such as future demand, reactions to new products, or brand loyalty to name only a few. Twitter analytics tailored for this task include community structure discovery algor

FAQs

What signals from Twitter are used to estimate account personality?

The approach uses structural and functional Twitter attributes, including original multimedia posting, retweeting patterns, and hashtag emotional polarity, along with reactions tied to political, social, or historical events.

How does the proposed intelligent agent generate MBTI estimates?

It employs a stochastic account visiting policy with preferential attachment, an optional evolving forgetting factor to penalize frequently appearing vertices, and a graph neural network to output MBTI estimates from the constructed graph.

Which evaluation aspects show the proposed method’s performance?

Results are assessed using criteria such as community size distribution and hashtag coherency, indicating superior performance for the proposed heuristic.

Sep 17, 2026

Uploaded by Skyler